To see the other types of publications on this topic, follow the link: Time series outlier detection.
Dissertations / Theses on the topic 'Time series outlier detection'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'Time series outlier detection.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Sedman, Robin. "Online Outlier Detection in Financial Time Series." Thesis, KTH, Matematisk statistik, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-228069.
Full textAbstract:
In this Master’s thesis, different models for outlier detection in financial time series are examined. The financial time series are price series such as index prices or asset prices. Outliers are, in this thesis, defined as extreme and false points, but this definition is also investigated and revised. Two different time series models are examined: an autoregressive (AR) and a generalized autoregressive conditional heteroskedastic (GARCH) time series model, as well as one test statistic method based on the GARCH model. Additionally, a nonparametric model is examined, which utilizes kernel density estimation in order to detect outliers. The models are evaluated by how well they detect outliers and how often they misclassify inliers as well as the run time of the models. It is found that all the models performs approximately equally good, on the data sets used in thesis and the simulations done, in terms of how well the methods find outliers, apart from the test static method which performs worse than the others. Furthermore it is found that definition of an outlier is very crucial to how well a model detects the outliers. For the application of this thesis, the run time is an important aspect, and with this in mind an autoregressive model with a Student’s t-noise distribution is found to be the best one, both with respect to how well it detects outliers, misclassify inliers and run time of the model.I detta examensarbete undersöks olika modeller för outlierdetektering i finansiella tidsserier. De finansiella tidsserierna är prisserier som indexpriser eller tillgångspriser. Outliers är i detta examensarbete definierade som extrema och falska punkter, men denna definition undersöks och revideras också. Två olika tidsseriemodeller undersöks: en autoregressiv (AR) och en generel au-toregressiv betingad heteroskedasticitet1 (GARCH) tidsseriemodell, samt en hypotesprövning2 baserad på GARCH-modellen. Dessutom undersöks en icke-parametrisk modell, vilken använder sig utav uppskattning av täthetsfunktionen med hjälp av kärnfunktioner3 för att detektera out-liers. Modellerna utvärderas utifrån hur väl de upptäcker outliers, hur ofta de kategoriserar icke-outliers som outliers samt modellens körtid. Det är konstaterat att alla modeller ungefär presterar lika bra, baserat på den data som används och de simuleringar som gjorts, i form av hur väl outliers är detekterade, förutom metoden baserad på hypotesprövning som fungerar sämre än de andra. Vidare är det uppenbart att definitionen av en outlier är väldigt avgörande för hur bra en modell detekterar outliers. För tillämpningen av detta examensarbete, så är körtid en viktig faktor, och med detta i åtanke är en autoregressiv modell med Students t-brusfördelning funnen att vara den bästa modellen, både med avseende på hur väl den detekterar outliers, felaktigt detekterar inliers som outliers och modellens körtid.
2
Wang, Dan Tong. "Outlier detection with data stream mining approach in high-dimenional time series data." Thesis, University of Macau, 2017. http://umaclib3.umac.mo/record=b3691091.
Full text
3
Bergamelli, Michele. "Structural breaks and outliers detection in time-series econometrics : methods and applications." Thesis, City University London, 2015. http://openaccess.city.ac.uk/14868/.
Full textAbstract:
This thesis contributes to the econometric literature on structural breaks analysis and outliers detection in parametric linear models. The focus is on the development of new econometric tools as well as on the analysis of novel but largely unexplored approaches. The econometric methods under analysis are illustrated using macroeconomic and financial relationships. The thesis is organised in three main chapters. In Chapter 2, we consider two novel methods to detect multiple structural breaks affecting the deterministic component of a linear system. The first is an extension of the dummy saturation method whereas the second method deals with a sequential bootstrapping procedure based on the sup-F statistic. Through an extensive Monte Carlo exercise, we explore the ability of the two approaches to detect the correct number and the correct location of the breaks. Additionally, we illustrate how to apply empirically the two procedures by investigating the stability of the Fisher relationship in the United States. In Chapter 3, we consider testing for multiple structural breaks in the vector error correction framework. First, we study the role of weak exogeneity when testing for structural breaks in the cointegrating matrix. Second, we extend the existing likelihood ratio test of Hansen (2003) to the case of unknown break dates through the specification of a minimum p-value statistic with critical values approximated by bootstrapping. Monte Carlo simulations show that the proposed statistic has good finite sample properties whilst three small empirical applications illustrate how the minimum p-value statistic can be used in practice. In Chapter 4, we tackle the purchasing power parity puzzle developing a robust estimator for the half-life of the real exchange rate. Specifically, we propose to identify outlying observations by means of a dummy saturation type algorithm designed for ARMA processes which enables to detect additional and innovative outliers as well as level shifts. An empirical application involving US dollar real exchange rates shows that the estimated half-lives are considerably shorter when outlying observations are correctly modelled, therefore shedding some light on the purchasing power parity puzzle.
4
Eghbalian, Amirmohammad. "Data mining techniques for modeling the operating behaviors of smart building control valve systems." Thesis, Blekinge Tekniska Högskola, Institutionen för datavetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:bth-20102.
Full textAbstract:
Background. One of the challenges about smart control valves system is processing and analyzing sensors data to extract useful information. These types of information can be used to detect the deviating behaviors which can be an indication of faults and issues in the system. Outlier detection is a process in which we try to find these deviating behaviors that occur in the system.Objectives. First, perform a literature review to get an insight about the machine learning (ML) and data mining (DM) techniques that can be applied to extract patternfrom time-series data. Next, model the operating behaviors of the control valve system using appropriate machine learning and data mining techniques. Finally,evaluate the proposed behavioral models on real world data.Methods. To have a better understanding of the different ML and DM techniques for extracting patterns from time-series data and fault detection and diagnosis of building systems, literature review is conducted. Later on, an unsupervised learning approach is proposed for modeling the typical operating behaviors and detecting the deviating operating behaviors of the control valve system. Additionally, the proposed method provides supplementary information for domain experts to help them in their analysis.Results. The outcome from modeling and monitoring the operating behaviors ofthe control valve system are analyzed. The evaluation of the results by the domain experts indicates that the method is capable of detecting deviating or unseen operating behaviors of the system. Moreover, the proposed method provides additional useful information to have a better understanding of the obtained results.Conclusions. The main goal in this study was achieved by proposing a method that can model the typical operating behaviors of the control valve system. The generated model can be used to monitor the newly arrived daily measurements and detect the deviating or unseen operating behaviors of the control valve system. Also, it provides supplementary information that can help domain experts to facilitate and reduce the time of analysis.
5
Slávik, Ľuboš. "Dynamická faktorová analýza časových řad." Master's thesis, Vysoké učení technické v Brně. Fakulta strojního inženýrství, 2021. http://www.nusl.cz/ntk/nusl-445469.
Full textAbstract:
Táto diplomová práca sa zaoberá novým prístupom k zhlukovaniu časových rád na základe dynamického faktorového modelu. Dynamický faktorový model je technika redukujúca dimenziu a rozširuje klasickú faktorovú analýzu o požiadavku autokorelačnej štruktúry latentných faktorov. Parametre modelu sa odhadujú pomocou EM algoritmu za použitia Kalmanovho filtra a vyhladzovača a taktiež sú aplikované nevyhnutné podmienky na model, aby sa stal identifikovateľným. Po tom, ako je v práci predstavený teoretický koncept prístupu, dynamický faktorový model je aplikovaný na skutočné pozorované časové rady a práca skúma jeho správanie a vlastnosti na jednomesačných meteorologických dátach požiarneho indexu (Fire Weather Index) na 108 požiarnych staniciach umiestnených v Britskej Kolumbii. Postup výpočtu modelu odhadne záťažovú maticu (loadings matrix) spolu so zodpovedajúcim malým počtom latentných faktorov a kovariačnou maticou modelovaných časových rád. Diplomová práca aplikuje k-means zhlukovanie na výslednú záťažovú maticu a ponúka rozdelenie meteorologických staníc do zhlukov založené na redukovanej dimenzionalite pôvodných dát. Vďaka odhadnutým priemerom zhlukov a odhadnutým latentným faktorom je možné získať aj priemerné trendy každého zhluku. Následne sú dosiahnuté výsledky porovnané s výsledkami získanými na dátach z rovnakých staníc avšak iného mesiaca, aby sa stanovila stabilita zhlukovania. Práca sa taktiež zaoberá efektom varimax rotácie záťažovej matice. Diplomová práca naviac navrhuje metódu detekovania odľahlých časových rád založenú na odhadnutej kovariačnej matici modelu a rozoberá dôsledky odľahlých hodnôt na odhanutý model.
6
Allab, Nedjmeddine. "Détection d'anomalies et de ruptures dans les séries temporelles. Applications à la gestion de production de l'électricité." Thesis, Paris 6, 2016. http://www.theses.fr/2016PA066658.
Full textAbstract:
Continental est l'outil de référence utilisé par EDF pour la gestion d'électricité à long terme. il permet d'élaborer la stratégie d'exploitation du parc constitué de centrales réparties sur toute l'europe. l'outil simule sur chaque zone et chaque scénario plusieurs variables telles que la demande d'électricité, la quantité générée ainsi que les coûts associés. nos travaux de thèse ont pour objectif de fournir des méthodes d'analyse de ces données de production afin de faciliter leur étude et leur synthèse. nous récoltons un ensemble de problématiques auprès des utilisateurs de continental que nous tentons de résoudre à l'aide des technique de détection d'anomalies et de ruptures dans les séries temporellesContinental is the main tool that edf uses for the long-term management of electricity. It elaborates the strategy exploitation of the electrical parc made up by power plants distributed all over europe. the tool simulates for each zone and each scenario several variables, such as the electricity demand, the generated quantity as well as the related costs. our works aim to provide methods to analyse the data of electricity production in order to ease their discovery and synthesis. we get a set of problmatics from the users of continental that we tent to solve through techniques of outliers and changepoints detection in time series
7
Ribeiro, Joana Patrícia Bordonhos. "Outlier identification in multivariate time series." Master's thesis, Universidade de Aveiro, 2017. http://hdl.handle.net/10773/22200.
Full textAbstract:
Mestrado em Matemática e AplicaçõesCom o desenvolvimento tecnológico, existe uma cada vez maior disponibilidade de dados. Geralmente representativos de situações do dia-a-dia, a existência de grandes quantidades de informação tem o seu interesse quando permite a extração de valor para o mercado. Além disso, surge importância em analisar não só os valores disponíveis mas também a sua associação com o tempo. A existência de valores anormais é inevitável. Geralmente denotados como outliers, a procura por estes valores é realizada comummente com o intuito de fazer a sua exclusão do estudo. No entanto, os outliers representam muitas vezes um objetivo de estudo. Por exemplo, no caso de deteção de fraudes bancárias ou no diagnóstico de doenças, o objetivo central é identificar situações anormais. Ao longo desta dissertação é apresentada uma metodologia que permite detetar outliers em séries temporais multivariadas, após aplicação de métodos de classificação. A abordagem escolhida é depois aplicada a um conjunto de dados real, representativo do funcionamento de caldeiras. O principal objetivo é identificar as suas falhas. Consequentemente, pretende-se melhorar os componentes do equipamento e portanto diminuir as suas falhas. Os algoritmos implementados permitem identificar não só as falhas do aparelho mas também o seu funcionamento normal. Pretende-se que as metodologias escolhidas sejam também aplicadas nos aparelhos futuros, permitindo melhorar a identificação em tempo real das falhas.
With the technological development, there is an increasing availability of data. Usually representative of day-to-day actions, the existence of large amounts of information has its own interest when it allows to extract value to the market. In addition, it is important to analyze not only the available values but also their association with time. The existence of abnormal values is inevitable. Usually denoted as outliers, the search for these values is commonly made in order to exclude them from the study. However, outliers often represent a goal of study. For example, in the case of bank fraud detection or disease diagnosis, the central objective is to identify the abnormal situations. Throughout this dissertation we present a methodology that allows the detection of outliers in multivariate time series, after application of classification methods. The chosen approach is then applied to a real data set, representative of boiler operation. The main goal is to identify faults. It is intended to improve boiler components and, hence, reduce the faults. The implemented algorithms allow to identify not only the boiler faults but also their normal operation cycles. We aim that the chosen methodologies will also be applied in future devices, allowing to improve real-time fault identification.
8
Åkerström, Emelie. "Real-time Outlier Detection using Unbounded Data Streaming and Machine Learning." Thesis, Luleå tekniska universitet, Datavetenskap, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-80044.
Full textAbstract:
Accelerated advancements in technology, the Internet of Things, and cloud computing have spurred an emergence of unstructured data that is contributing to rapid growth in data volumes. No human can manage to keep up with monitoring and analyzing these unbounded data streams and thus predictive and analytic tools are needed. By leveraging machine learning this data can be converted into insights which are enabling datadriven decisions that can drastically accelerate innovation, improve user experience, and drive operational efficiency. The purpose of this thesis is to design and implement a system for real-time outlier detection using unbounded data streams and machine learning. Traditionally, this is accomplished by using alarm-thresholds on important system metrics. Yet, a static threshold cannot account for changes in trends and seasonality, changes in the system, or an increased system load. Thus, the intention is to leverage machine learning to instead look for deviations in the behavior of the data not caused by natural changes but by malfunctions. The use-case driving the thesis forward is real-time outlier detection in a Content Delivery Network (CDN). The input data includes Http-error messages received by clients, and contextual information like region, cache domains, and error codes, to provide tailormade predictions accounting for the trends in the data. The outlier detection system consists of a data collection pipeline leveraging the technique of stream processing, a MiniBatchKMeans clustering model that provides online clustering of incoming data according to their similar characteristics, and an LSTM AutoEncoder that accounts for temporal nature of the data and detects outlier data points in the clusters. An important finding is that an outlier is defined as an abnormal amount of outlier data points all originating from the same cluster, not a single outlier data point. Thus, the alerting system will be implementing an outlier percentage threshold. The experimental results show that an outlier is detected within one minute from a cache break-down. This triggers an alert to the system owners, containing graphs of the clustered data to narrow down the search area of the cause to enable preventive action towards the prominent incident. Further results show that within 2 minutes from fixing the cause the system will provide feedback that the actions taken were successful. Considering the real-time requirements of the CDN environment, it is concluded that the short delay for detection is indeed real-time. Proving that machine learning is indeed able to detect outliers in unbounded data streams in a real-time manner. Further analysis shows that the system is more accurate during peakhours when more data is in circulation than during none peak-hours, despite the temporal LSTM layers. Presumably, an effect from the model needing to train on more data to better account for seasonality and trends. Future work necessary to put the outlier detection system in production thus includes more training to improve accuracy and correctness. Furthermore, one could consider implementing necessary functionality for a production environment and possibly adding enhancing features that can automatically avert incidents detected and handle the causes of them.
9
Sánchez, Enríquez Heider Ysaías. "Anomaly detection in streaming multivariate time series." Tesis, Universidad de Chile, 2017. http://repositorio.uchile.cl/handle/2250/149078.
Full textAbstract:
Doctor en Ciencias, Mención ComputaciónEste trabajo de tesis presenta soluciones para al problema de detección de anomalı́as en flujo de datos multivariantes. Dado una subsequencia de serie temporal (una pequeña parte de la serie original) como entrada, uno quiere conocer si este corresponde a una observación normal o es una anomalı́a, con respecto a la información histórica. Pueden surgir dificultades debido principalmente a que los tipos de anomalı́a son desconocidos. Además, la detección se convierte en una tarea costosa debido a la gran cantidad de datos y a la existencia de variables de dominios heterogéneos. En este contexto, se propone un enfoque de detección de anomalı́as basado en Discord Discovery, que asocia la anomalı́a con la subsecuencia más inusual utilizando medidas de similitud. Tı́picamente, los métodos de reducción de la dimensionalidad y de indexación son elaborados para restringir el problema resolviéndolo eficientemente. Adicionalmente, se propone técnicas para generar modelos representativos y consisos a partir de los datos crudos con el fin de encontrar los patrones inusuales. Estas técnicas también mejoran la eficiencia en la búsqueda mediante la reducción de la dimensionalidad. Se aborda las series multivariantes usando técnicas de representación sobre subsequencias no- normalizadas, y se propone nuevas técnicas de discord discovery basados en ı́ndices métricos. El enfoque propuesto es comparado con técnicas del estado del arte. Los resultados ex- perimentales demuestran que aplicando la transformación de translación y representación de series temporales pueden contribuir a mejorar la eficacia en la detección. Además, los métodos de indexación métrica y las heurı́sticas de discord discovery pueden resolver eficien- temente la detección de anomalı́as en modo offline y online en flujos de series temporales multivariantes.
Este trabajo ha sido financiado por beca CONICYT - CHILE / Doctorado para Extranjeros, y apoyada parcialmente por el Proyecto FONDEF D09I1185 y el Programa de Becas de NIC Chile
10
Tinawi, Ihssan. "Machine learning for time series anomaly detection." Thesis, Massachusetts Institute of Technology, 2019. https://hdl.handle.net/1721.1/123129.
Full textAbstract:
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2019
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (page 55).
In this thesis, I explored machine learning and other statistical techniques for anomaly detection on time series data obtained from Internet-of-Things sensors. The data, obtained from satellite telemetry signals, were used to train models to forecast a signal based on its historic patterns. When the prediction passed a dynamic error threshold, then that point was flagged as anomalous. I used multiple models such as Long Short-Term Memory (LSTM), autoregression, Multi-Layer Perceptron, and Encoder-Decoder LSTM. I used the "Detecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholding" paper as a basis for my analysis, and was able to beat their performance on anomaly detection by obtaining an F0.5 score of 76%, an improvement over their 69% score.
by Ihssan Tinawi.
M. Eng.
M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science
11
Sperl, Ryan E. "Hierarchical Anomaly Detection for Time Series Data." Wright State University / OhioLINK, 2020. http://rave.ohiolink.edu/etdc/view?acc_num=wright1590709752916657.
Full text
12
Roth, Jennifer M. "A Time Series Approach to Removing Outlying Data Points from Bluetooth Vehicle Speed Data." University of Akron / OhioLINK, 2010. http://rave.ohiolink.edu/etdc/view?acc_num=akron1289758088.
Full text
13
Ji, Kang Hyeun. "Transient signal detection using GPS position time series." Thesis, Massachusetts Institute of Technology, 2011. http://hdl.handle.net/1721.1/69466.
Full textAbstract:
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Earth, Atmospheric, and Planetary Sciences, 2011.Cataloged from PDF version of thesis.
Includes bibliographical references (p. 229-243).
Continuously operating Global Positioning System (GPS) networks record station position changes with millimeter-level accuracy and have revealed transient deformations on various spatial and temporal scales. However, the transient deformation may not be easily identified from the position time series because of low signal-to-noise ratios (SNR), correlated noise in space and time and large number of sites in a network. As a systematic detection method, we use state estimation based on Kalman filtering and principal component analysis (PCA). State estimation improves the SNR in the time domain by estimating secular and transient motions and reducing the level of both white and colored noise. PCA improves the SNR in space domain by accounting for the coherence of transient motions between nearby sites. Synthetic tests show that the method is capable of detecting transient signals embedded in noisy data but complex signals (e.g., large-scale signals in space and time, multiple and/or propagating signals) are difficult to detect and interpret. We demonstrate the detection capability with two known signals in the Los Angeles basin, California: far-field coseismic offsets associated with the 1999 Hector Mine earthquake and locally-observed hydrologic deformation due to heavy rainfall in winter 2004-2005 in San Gabriel Valley. We applied the method to the daily GPS data from the Plate Boundary Observatory (PBO) network in Alaska and in the Washington State section of the Cascadia subduction zone. We have detected a transient signal whose maximum displacement is -9 mm in the horizontal and -41 mm in the vertical at Akutan volcano, Alaska, during the first half of 2008 and two previously unrecognized small slow slip events with average surface displacements less than 2 mm, which was thought to be below current GPS resolution. The detection method improves the SNR and therefore provides higher resolution for detecting weak transient signals, and it can be used as a routine monitoring system.
by Kang Hyeun Ji.
Ph.D.
14
Lindroth, Henriksson Amelia. "Unsupervised Anomaly Detection on Time Series Data: An Implementation on Electricity Consumption Series." Thesis, KTH, Matematisk statistik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-301731.
Full textAbstract:
Digitization of the energy industry, introduction of smart grids and increasing regulation of electricity consumption metering have resulted in vast amounts of electricity data. This data presents a unique opportunity to understand the electricity usage and to make it more efficient, reducing electricity consumption and carbon emissions. An important initial step in analyzing the data is to identify anomalies. In this thesis the problem of anomaly detection in electricity consumption series is addressed using four machine learning methods: density based spatial clustering for applications with noise (DBSCAN), local outlier factor (LOF), isolation forest (iForest) and one-class support vector machine (OC-SVM). In order to evaluate the methods synthetic anomalies were introduced to the electricity consumption series and the methods were then evaluated for the two anomaly types point anomaly and collective anomaly. In addition to electricity consumption data, features describing the prior consumption, outdoor temperature and date-time properties were included in the models. Results indicate that the addition of the temperature feature and the lag features generally impaired anomaly detection performance, while the inclusion of date-time features improved it. Of the four methods, OC-SVM was found to perform the best at detecting point anomalies, while LOF performed the best at detecting collective anomalies. In an attempt to improve the models' detection power the electricity consumption series were de-trended and de-seasonalized and the same experiments were carried out. The models did not perform better on the decomposed series than on the non-decomposed.Digitaliseringen av elbranschen, införandet av smarta nät samt ökad reglering av elmätning har resulterat i stora mängder eldata. Denna data skapar en unik möjlighet att analysera och förstå fastigheters elförbrukning för att kunna effektivisera den. Ett viktigt inledande steg i analysen av denna data är att identifiera möjliga anomalier. I denna uppsats testas fyra olika maskininlärningsmetoder för detektering av anomalier i elförbrukningsserier: densitetsbaserad spatiell klustring för applikationer med brus (DBSCAN), lokal avvikelse-faktor (LOF), isoleringsskog (iForest) och en-klass stödvektormaskin (OC-SVM). För att kunna utvärdera metoderna infördes syntetiska anomalier i elförbrukningsserierna och de fyra metoderna utvärderades därefter för de två anomalityperna punktanomali och gruppanomali. Utöver elförbrukningsdatan inkluderades även variabler som beskriver tidigare elförbrukning, utomhustemperatur och tidsegenskaper i modellerna. Resultaten tyder på att tillägget av temperaturvariabeln och lag-variablerna i allmänhet försämrade modellernas prestanda, medan införandet av tidsvariablerna förbättrade den. Av de fyra metoderna visade sig OC-SVM vara bäst på att detektera punktanomalier medan LOF var bäst på att detektera gruppanomalier. I ett försök att förbättra modellernas detekteringsförmåga utfördes samma experiment efter att elförbrukningsserierna trend- och säsongsrensats. Modellerna presterade inte bättre på de rensade serierna än på de icke-rensade.
15
Yacob, Andreas, and Olof Nilsson. "Non-parametric anomaly detection in sentiment time series data." Thesis, Uppsala universitet, Institutionen för teknikvetenskaper, 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-251645.
Full textAbstract:
The importance of finding extreme events or unexpected patterns has increased over the last two decades, mainly due rapid advancements in technology. These events or patterns are referred to as anomalies. This thesis focuses on detecting anomalies in form of sudden peaks occurring in time series generated from online text analysis in Gavagai’s live environment. To our knowledge there exist a limited number of sequential peak detection models applicable in this domain. We introduce a novel technique using the Local Outlier Factor model as well as a model built on simple linear regression with a Bayesian error function, both operating in real-time. We also study a model based on linear Poisson regression. With the constraint from Gavagai that the models should be easy to setup for different targets, it requires them to be non-parametric. The Local Outlier Factor model and the simple linear regression model show promising results comparing them to Gavagai’s current working model. All models were tested on 3 datasets representing 3 different sentiment targets; positivity, negativity and frequency. Not only do our models superiorly succeed to detect the anomalies, but also they do so with fixed parameters independent of target looked at. This means that our models have lower error rate even though they are non-parametric constructed, compared to Gavagai’s current model that requires tuning per target of interest to operate with sufficient accuracy.
16
Barham, S. Y. "Time series analysis in the detection of breast cancer." Thesis, Bucks New University, 1987. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.384665.
Full text
17
Vendramin, Nicoló. "Unsupervised Anomaly Detection on Multi-Process Event Time Series." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-254885.
Full textAbstract:
Establishing whether the observed data are anomalous or not is an important task that has been widely investigated in literature, and it becomes an even more complex problem if combined with high dimensional representations and multiple sources independently generating the patterns to be analyzed. The work presented in this master thesis employs a data-driven pipeline for the definition of a recurrent auto-encoder architecture to analyze, in an unsupervised fashion, high-dimensional event time-series generated by multiple and variable processes interacting with a system. Facing the above mentioned problem the work investigates whether it is possible or not to use a single model to analyze patterns produced by different sources. The analysis of log files that record events of interaction between users and the radio network infrastructure is employed as realworld case-study for the given problem. The investigation aims to verify the performances of a single machine learning model applied to the learning of multiple patterns developed through time by distinct sources. The work proposes a pipeline, to deal with the complex representation of the data source and the definition and tuning of the anomaly detection model, that is based on no domain-specific knowledge and can thus be adapted to different problem settings. The model has been implemented in four different variants that have been evaluated over both normal and anomalous data, gathered partially from real network cells and partially from the simulation of anomalous behaviours. The empirical results show the applicability of the model for the detection of anomalous sequences and events in the described conditions, with scores reaching above 80% in terms of F1-score, and varying depending on the specific threshold setting. In addition, their deeper interpretation gives insights about the difference between the variants of the model and thus, their limitations and strong points.Att fastställa huruvida observerade data är avvikande eller inte är en viktig uppgift som har studerats ingående i litteraturen och problemet blir ännu mer komplext, om detta kombineras med högdimensionella representationer och flera källor som oberoende genererar de mönster som ska analyseras. Arbetet som presenteras i denna uppsats använder en data-driven pipeline för definitionen av en återkommande auto-encoderarkitektur för att analysera, på ett oövervakat sätt, högdimensionella händelsetidsserier som genereras av flera och variabla processer som interagerar med ett system. Mot bakgrund av ovanstående problem undersöker arbetet om det är möjligt eller inte att använda en enda modell för att analysera mönster som producerats av olika källor. Analys av loggfiler som registrerar händelser av interaktion mellan användare och radionätverksinfrastruktur används som en fallstudie för det angivna problemet. Undersökningen syftar till att verifiera prestandan hos en enda maskininlärningsmodell som tillämpas för inlärning av flera mönster som utvecklats över tid från olika källor. Arbetet föreslår en pipeline för att hantera den komplexa representationen hos datakällorna och definitionen och avstämningen av anomalidetektionsmodellen, som inte är baserad på domänspecifik kunskap och därför kan anpassas till olika probleminställningar. Modellen har implementerats i fyra olika varianter som har utvärderats med avseende på både normala och avvikande data, som delvis har samlats in från verkliga nätverksceller och delvis från simulering av avvikande beteenden. De empiriska resultaten visar modellens tillämplighet för detektering av avvikande sekvenser och händelser i det föreslagna ramverket, med F1-score över 80%, varierande beroende på den specifika tröskelinställningen. Dessutom ger deras djupare tolkning insikter om skillnaden mellan olika varianter av modellen och därmed deras begränsningar och styrkor.
18
Granlund, Oskar. "Unsupervised anomaly detection on log-based time series data." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-265534.
Full textAbstract:
Due to a constant increase in the number of connected devices and there is an increased demand for confidentiality, availability, and integrity on applications. This thesis was focused on unsupervised anomaly detection in data centers. It evaluates how suitable open source state-of-the-art solutions are at finding abnormal trends and patterns in log-based data streams. The methods used in this work are Principal component analysis (PCA), LogCluster, and Hierarchical temporal memory (HTM). They were evaluated using F-score on a real data set from an Apache access log. The data set was carefully chosen to represent a normal state in which close to no anomalous events occurred. Af- terward, 0.5% of the data points were transformed into anomalous data points, calculated based on the average frequency of log events matching a certain pattern. PCA showed the best performance with an F-score ranging from 0.4 - 0.56. The second best method was LogCluster but the HTM methods did not show adequate results. The result showed that PCA can find approximately 50% of the injected anomalies, this can be used to improve the confidentiality, integrity and availability of applications.Eftersom antalet uppkopplade enheter ständigt har ökat och kravet på tillgänglighet, äkthet och integritet hos applikationer är höga så har den här uppsatsen fokuserat på oövervakad anomalidetektering i datacenter. Den utvärderar hur lämpliga öppna och moderna anomalidetekteringsmetoder är för att hitta avvikande mönster och trender på logbaserade dataströmmar. Metoderna använda i det här projektet är Principalkomponentanalys, LogCluster och Hierarkisk temporärt minne. De är utvärderade med F-score på en datamängd från en Apache-accesslogg tagen från en produktionsmiljö. Datan var utvald för att reprensentera ett normalt tillstånd där få eller inga onormala händelser förekom. 0.5% av datapunkterna transformerades till anomalier, baserat på den genomsnittliga förekomsten av varje logsekvens som matchar ett visst mönster. Principalkomponentanalys visade de bästa resultaten med ett F-score från 0.4 till 0.56. Näst bäst var LogCluster, de två metoderna baserade på hierarkiskt temporärt minne visade inte alls bra resultat. Resultaten visade att PCA kan hitta ca 50% av de injecerade anomalierna vilket kan användas för att förbättra konfidentialitet, tillgänglighet och integriteten hos applikationer.
19
Wolpher, Maxim. "Anomaly Detection in Unstructured Time Series Datausing an LSTM Autoencoder." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-231368.
Full textAbstract:
An exploration of anomaly detection. Much work has been done on the topic of anomalyd etection, but what seems to be lacking is a dive into anomaly detection of unstructuredand unlabeled data. This thesis aims to determine the efctiveness of combining recurrentneural networks with autoencoder structures for sequential anomaly detection. The use of an LSTM autoencoder will be detailed, but along the way there will also be backgroundon time-independent anomaly detection using Isolation Forests and Replicator Neural Networks on the benchmark DARPA dataset. The empirical results in this thesis show that Isolation Forests and Replicator Neural Networks both reach an F1-score of 0.98. The RNN reached a ROC AUC score of 0.90 while the Isolation Forest reached a ROC AUC of 0.99. The results for the LSTM Autoencoder show that with 137 features extracted from the unstructured data, it can reach an F1 score of 0.8 and a ROC AUC score of 0.86En undersökning av anomalitetsdetektering. Mycket arbete har gjorts inom ämnet anomalitetsdetektering, men det som verkar saknas är en fördjupning i anomalitetsdetektering av ostrukturerad och omärktdata. Denna avhandling syftar till att bestämma effektiviteten av att kombinera Recurrent Neural Networks med Autoencoder strukturer för sekventiell anomalitetsdetektion. Användningen av en LSTM autoencoder kommeratt beskrivas i detalj, men bakgrund till tidsoberoende anomalitetsdetektering med hjälp av Isolation Forests och Replicator Neural Networks på referens DARPA dataset kommer också att täckas. De empiriska resultaten i denna avhandling visar att Isolation Forestsoch Replicator Neural Networks (RNN) båda når en F1-score på 0,98. RNN nådde en ROC AUC-score på 0,90 medan Isolation Forest nådde en ROC-AUC på 0,99. Resultaten för LSTM Autoencoder visar att med 137 features extraherade från ostrukturerad data kan den nå en F1-score på 0,80 och en ROC AUC-score på 0,86.
20
Haddad, Josef, and Carl Piehl. "Unsupervised anomaly detection in time series with recurrent neural networks." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-259655.
Full textAbstract:
Artificial neural networks (ANN) have been successfully applied to a wide range of problems. However, most of the ANN-based models do not attempt to model the brain in detail, but there are still some models that do. An example of a biologically constrained ANN is Hierarchical Temporal Memory (HTM). This study applies HTM and Long Short-Term Memory (LSTM) to anomaly detection problems in time series in order to compare their performance for this task. The shape of the anomalies are restricted to point anomalies and the time series are univariate. Pre-existing implementations that utilise these networks for unsupervised anomaly detection in time series are used in this study. We primarily use our own synthetic data sets in order to discover the networks’ robustness to noise and how they compare to each other regarding different characteristics in the time series. Our results shows that both networks can handle noisy time series and the difference in performance regarding noise robustness is not significant for the time series used in the study. LSTM outperforms HTM in detecting point anomalies on our synthetic time series with sine curve trend but a conclusion about the overall best performing network among these two remains inconclusive.Artificiella neurala nätverk (ANN) har tillämpats på många problem. Däremot försöker inte de flesta ANN-modeller efterlikna hjärnan i detalj. Ett exempel på ett ANN som är begränsat till att efterlikna hjärnan är Hierarchical Temporal Memory (HTM). Denna studie tillämpar HTM och Long Short-Term Memory (LSTM) på avvikelsedetektionsproblem i tidsserier för att undersöka vilka styrkor och svagheter de har för detta problem. Avvikelserna i denna studie är begränsade till punktavvikelser och tidsserierna är i endast en variabel. Redan existerande implementationer som utnyttjar dessa nätverk för oövervakad avvikelsedetektionsproblem i tidsserier används i denna studie. Vi använder främst våra egna syntetiska tidsserier för att undersöka hur nätverken hanterar brus och hur de hanterar olika egenskaper som en tidsserie kan ha. Våra resultat visar att båda nätverken kan hantera brus och prestationsskillnaden rörande brusrobusthet var inte tillräckligt stor för att urskilja modellerna. LSTM presterade bättre än HTM på att upptäcka punktavvikelser i våra syntetiska tidsserier som följer en sinuskurva men en slutsats angående vilket nätverk som presterar bäst överlag är fortfarande oavgjord.
21
Berenji, Ardestani Sarah. "Time Series Anomaly Detection and Uncertainty Estimation using LSTM Autoencoders." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-281354.
Full textAbstract:
The goal of this thesis is to implement an anomaly detection tool using LSTM autoencoder and apply a novel method for uncertainty estimation using Bayesian NeuralNetworks (BNNs) based on a paper from Uber research group [1]. Having a reliable anomaly detection tool and accurate uncertainty estimation is critical in many fields. At Telia, such a tool can be used in many different data domains like device logs to detect abnormal behaviours. Our method uses an autoencoder to extract important features and learn the encoded representation of the time series. This approach helps to capture testing data points coming from a different population. We then train a prediction model based on this encoder’s representation of data. An uncertainty estimation algorithm is used to estimate the model’s uncertainty, which breaks it down to three different sources: model uncertainty, model misspecification, and inherent noise. To get the first two, a Monte Carlo dropout approach is used which is simple to implement and easy to scale. For the third part, a bootstrap approach that estimates the noise level via the residual sum of squares on validation data is used. As a result, we could see that our proposed model can make a better prediction in comparison to our benchmarks. Although the difference is not big, yet it shows that making prediction based on encoding representation is more accurate. The anomaly detection results based on these predictions also show that our proposed model has a better performance than the benchmarks. This means that using autoencoder can improve both prediction and anomaly detection tasks. Additionally, we conclude that using deep neutral networks would show bigger improvement if the data has more complexity.Målet med den här uppsatsen är att implentera ett verktyg för anomaliupptäckande med hjälp av LSTM autoencoders och applicera en ny metod för osäkerhetsestimering med hjälp av Bayesian Neural Networks (BNN) baserat på en artikel från Uber research group [1]. Pålitliga verktyg för att upptäcka anomalier och att göra precisa osäkerhetsestimeringar är kritiskt i många fält. På Telia kan ett sådant verktyg användas för många olika datadomäner, som i enhetsloggar för att upptäcka abnormalt beteende. Vår metod använder en autoencoder för att extrahera viktiga egenskaper och lära sig den kodade representationen av tidsserierna. Detta tillvägagångssätt hjälper till med att ta in testdatapunker som kommer in från olika grundmängder. Sedan tränas en förutsägelsemodell baserad på encoderns representation av datan. För att uppskatta modellens osäkerhet används en uppskattningsalgoritm som delar upp osäkerheten till tre olika källor. Dessa tre källor är: modellosäkerhet, felspeciferad model, och naturligt brus. För att få de första två används en Monte Carlo dropout approach som är lätt att implementera och enkel att skala. För den tredje delen används en enkel anfallsvikel som uppskattar brusnivån med hjälp av felkvadratsumman av valideringsdatan. Som ett resultat kunde vi se att vår föreslagna model kan göra bättre förutsägelser än våra benchmarks. Även om skillnaden inte är stor så visar det att att använda autoencoderrepresentation för att göra förutsägelser är mer noggrant. Resulaten för anomaliupptäckanden baserat på dessa förutsägelser visar också att vår föreslagna modell har bättre prestanda än benchmarken. Det betyder att användning av autoencoders kan förbättra både förutsägelser och anomaliupptäckande. Utöver det kan vi dra slutsatsen att användning av djupa neurala nätverk skulle visa en större förbättring om datan hade mer komplexitet.
22
Thorén, Sofia, and Richard Sörberg. "Anomaly Detection in Signaling Data Streams : A Time-Series Approach." Thesis, Umeå universitet, Institutionen för matematik och matematisk statistik, 2017. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-139773.
Full textAbstract:
This work has aimed to develop a method which can be used in order to detect anomalies in signaling data streams at a telecommunication company. It has been done by modeling and evaluating three prediction models and two detection methods. The prediction models which have been implemented are Autoregressive Integrated Moving Average (ARIMA), Holt-Winters and a Benchmark model, furthermore have two detection methods been tested; Method 1 (M1), which is based on a robust evaluation of previous prediction errors and Method 2 (M2), which is based on the standard deviation in previous data. From the evaluation of the work, we could conclude that the best performing combination of prediction- and detection methods was achieved with a modified Benchmark model and M1- detection.
23
Du, Yang. "Comparison of change-point detection algorithms for vector time series." Thesis, Linköpings universitet, Statistik, 2010. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-59925.
Full textAbstract:
Change-point detection aims to reveal sudden changes in sequences of data. Special attention has been paid to the detection of abrupt level shifts, and applications of such techniques can be found in a great variety of fields, such as monitoring of climate change, examination of gene expressions and quality control in the manufacturing industry. In this work, we compared the performance of two methods representing frequentist and Bayesian approaches, respectively. The frequentist approach involved a preliminary search for level shifts using a tree algorithm followed by a dynamic programming algorithm for optimizing the locations and sizes of the level shifts. The Bayesian approach involved an MCMC (Markov chain Monte Carlo) implementation of a method originally proposed by Barry and Hartigan. The two approaches were implemented in R and extensive simulations were carried out to assess both their computational efficiency and ability to detect abrupt level shifts. Our study showed that the overall performance regarding the estimated location and size of change-points was comparable for the Bayesian and frequentist approach. However, the Bayesian approach performed better when the number of change-points was small; whereas the frequentist became stronger when the change-point proportion increased. The latter method was also better at detecting simultaneous change-points in vector time series. Theoretically, the Bayesian approach has a lower computational complexity than the frequentist approach, but suitable settings for the combined tree and dynamic programming can greatly reduce the processing time.
24
Santos, Rui Pedro Silvestre dos. "Time series morphological analysis applied to biomedical signals events detection." Master's thesis, Faculdade de Ciências e Tecnologia, 2011. http://hdl.handle.net/10362/10227.
Full textAbstract:
Dissertation submitted in the fufillment of the requirements for the Degree of Master in Biomedical EngineeringAutomated techniques for biosignal data acquisition and analysis have become increasingly powerful, particularly at the Biomedical Engineering research field. Nevertheless, it is verified the need to improve tools for signal pattern recognition and classification systems, in which the detection of specific events and the automatic signal segmentation are preliminary processing steps. The present dissertation introduces a signal-independent algorithm, which detects significant events in a biosignal. From a time series morphological analysis, the algorithm computes the instants when the most significant standard deviation discontinuities occur, segmenting the signal. An iterative optimization step is then applied. This assures that a minimal error is achieved when modeling these segments with polynomial regressions. The adjustment of a scale factor gives different detail levels of events detection. An accurate and objective algorithm performance evaluation procedure was designed. When applied on a set of synthetic signals, with known and quantitatively predefined events, an overall mean error of 20 samples between the detected and the actual events showed the high accuracy of the proposed algorithm. Its ability to perform the detection of signal activation onsets and transient waveshapes was also assessed, resulting in higher reliability than signal-specific standard methods. Some case studies, with signal processing requirements for which the developed algorithm can be suitably applied, were approached. The algorithm implementation in real-time, as part of an application developed during this research work, is also reported. The proposed algorithm detects significant signal events with accuracy and significant noise immunity. Its versatile design allows the application in different signals without previous knowledge on their statistical properties or specific preprocessing steps. It also brings added objectivity when compared with the exhaustive and time-consuming examiner analysis. The tool introduced in this dissertation represents a relevant contribution in events detection, a particularly important issue within the wide digital biosignal processing research field.
25
Mian, Ammar. "Contributions to SAR Image Time Series Analysis." Thesis, Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLC074/document.
Full textAbstract:
La télédétection par Radar à Synthèse d’Ouverture (RSO) offre une opportunité unique d’enregistrer, d’analyser et de prédire l’évolution de la surface de la Terre. La dernière décennie a permis l’avènement de nombreuses missions spatiales équipées de capteurs RSO (Sentinel-1, UAVSAR, TerraSAR X, etc.), ce qui a engendré une rapide amélioration des capacités d’acquisition d’images de la surface de la Terre. Le nombre croissant d’observations permet maintenant de construire des bases de données caractérisant l’évolution temporelle d’images, augmentant considérablement l’intérêt de l’analyse de séries temporelles pour caractériser des changements qui ont lieu à une échelle globale. Cependant, le développement de nouveaux algorithmes pour traiter ces données très volumineuses est un défi qui reste à relever. Dans ce contexte, l’objectif de cette thèse consiste ainsi à proposer et à développer des méthodologies relatives à la détection de changements dans les séries d’images ROS à très haute résolution spatiale.Le traitement de ces séries pose deux problèmes notables. En premier lieu, les méthodes d’analyse statistique performantes se basent souvent sur des données multivariées caractérisant, dans le cas des images RSO, une diversité polarimétrique, interférométrique, par exemple. Lorsque cette diversité n’est pas disponible et que les images RSO sont monocanal, de nouvelles méthodologies basées sur la décomposition en ondelettes ont été développées. Celles-ci permettent d’ajouter une diversité supplémentaire spectrale et angulaire représentant le comportement physique de rétrodiffusion des diffuseurs présents la scène de l’image. Dans un second temps, l’amélioration de la résolution spatiale sur les dernières générations de capteurs engendre une augmentation de l’hétérogénéité des données obtenues. Dans ce cas, l’hypothèse gaussienne, traditionnellement considérée pour développer les méthodologies standards de détection de changements, n’est plus valide. En conséquence, des méthodologies d’estimation robuste basée sur la famille des distributions elliptiques, mieux adaptée aux données, ont été développées.L’association de ces deux aspects a montré des résultats prometteurs pour la détection de changements.Le traitement de ces séries pose deux problèmes notables. En premier lieu, les méthodes d’analyse statistique performantes se basent souvent sur des données multivariées caractérisant, dans le cas des images RSO, une diversité polarimétrique ou interférométrique, par exemple. Lorsque cette diversité n’est pas disponible et que les images RSO sont monocanal, de nouvelles méthodologies basées sur la décomposition en ondelettes ont été développées. Celles-ci permettent d’ajouter une diversité spectrale et angulaire supplémentaire représentant le comportement physique de rétrodiffusion des diffuseurs présents la scène de l’image. Dans un second temps, l’amélioration de la résolution spatiale sur les dernières générations de capteurs engendre une augmentation de l’hétérogénéité des données obtenues. Dans ce cas, l’hypothèse gaussienne, traditionnellement considérée pour développer les méthodologies standards de détection de changements, n’est plus valide. En conséquence, des méthodologies d’estimation robuste basée sur la famille des distributions elliptiques, mieux adaptée aux données, ont été développées.L’association de ces deux aspects a montré des résultats prometteurs pour la détection de changementsRemote sensing data from Synthetic Aperture Radar (SAR) sensors offer a unique opportunity to record, to analyze, and to predict the evolution of the Earth. In the last decade, numerous satellite remote sensing missions have been launched (Sentinel-1, UAVSAR, TerraSAR X, etc.). This resulted in a dramatic improvement in the Earth image acquisition capability and accessibility. The growing number of observation systems allows now to build high temporal/spatial-resolution Earth surface images data-sets. This new scenario significantly raises the interest in time-series processing to monitor changes occurring over large areas. However, developing new algorithms to process such a huge volume of data represents a current challenge. In this context, the present thesis aims at developing methodologies for change detection in high-resolution SAR image time series.These series raise two notable challenges that have to be overcome:On the one hand, standard statistical methods rely on multivariate data to infer a result which is often superior to a monovariate approach. Such multivariate data is however not always available when it concerns SAR images. To tackle this issue, new methodologies based on wavelet decomposition theory have been developed to fetch information based on the physical behavior of the scatterers present in the scene.On the other hand, the improvement in resolution obtained from the latest generation of sensors comes with an increased heterogeneity of the data obtained. For this setup, the standard Gaussian assumption used to develop classic change detection methodologies is no longer valid. As a consequence, new robust methodologies have been developed considering the family of elliptical distributions which have been shown to better fit the observed data.The association of both aspects has shown promising results in change detection applications
26
Taillade, Thibault. "A new strategy for change detection in SAR time-series : application to target detection." Thesis, université Paris-Saclay, 2020. http://www.theses.fr/2020UPAST050.
Full textAbstract:
La détection de cibles telles que des navires ou des véhicules dans les images SAR (Synthetic Aperture radar) est un défi important pour la surveillance et la sécurité. Dans certains environnements tels que les zones urbaines, portuaires ou les forêts observées à basses fréquences radar, la détection de ces objets devient difficile en raison des propriétés de rétrodiffusion élevées de l'environnement. Pour résoudre ce problème, la détection de changement (CD) entre différentes images SAR permet de supprimer l'effet de l'environnement et ainsi une meilleur détection des cibles. Cependant, dans différents environnements à forte fréquentation, un chevauchement temporel des cibles peut se produire et génère une erreur d'interprétation possible car l'issue de la détection de changement repose sur une différence relative entre des objets de tailles ou de propriétés différentes. C'est un problème critique lorsque le but est de visualiser et d'obtenir le nombre d'objets à une acquisition donnée, dans les zones à fortes fréquentations comme les ports ou les zones urbaines. Idéalement, cette détection de changement devrait se réaliser entre une image constituée seulement de l'environnement et une image contenant les cibles d’intérêts. Grâce à l'accessibilité actuelle aux séries temporelles d'images SAR, nous proposons de générer une scène de référence (Frozen Background Image - FBR) qui n'est constituée que de l'environnement temporellement statique. La détection de changement entre une image SAR et cette image FBR vise donc a obtenir une map de détection des cibles éphémères présentes. Cette stratégie a été mise en œuvre pour la détection des navires en milieu portuaire et dans le contexte de véhicules cachés sous couvert forestierThe detection of targets such as ships or vehicles in SAR (Synthetic Aperture Radar) images is an essential challenge for surveillance and security purpose. In some environments such as urban areas, harbor areas or forest observed at low radar frequencies, detecting these objects becomes difficult due to the high backscattering properties of the surrounding background. To overcome this issue, change detection (CD) between SAR images enables to cancel the background and highlight successfully targets present within the scene. However, in several environments, a temporal overlapping of targets may occur and generates possible misinterpretation because the outcome relies on the relative change between objects of different sizes or properties. This is a critical issue when the purpose is to visualize and obtain the number of targets at a specific day in high attendance areas such as harbors or urban environments. Ideally, this change detection should occur between a target-free image and onewith possible objects of interest. With the current accessibility to SAR time-series, we propose to compute a frozen background reference (FBR) image that will consists only in the temporally static background. Performing change detection from this FBR image and any SAR image aim to highlight the presence of ephemeral targets. This strategy has been implemented for ship detection in harbor environment and in the context of vehicles hidden under foliage
27
Stafford, William B. "Sequential pattern detection and time series models for predicting IED attacks." Thesis, Monterey, Calif. : Naval Postgraduate School, 2009. http://edocs.nps.edu/npspubs/scholarly/theses/2009/Mar/09Mar%5FStafford.pdf.
Full textAbstract:
Thesis (M.S. in Information Technology Management)--Naval Postgraduate School, March 2009.Thesis Advisor(s): Kamel, Magdi. "March 2009." Description based on title screen as viewed on April 24, 2009. Author(s) subject terms: Sequential Pattern Detection, Time Series, Predicting IED Attacks, Data Mining. Includes bibliographical references (p. 77). Also available in print.
28
Liu, R. Q., and Ch Jacobi. "Piecewise linear trend detection in mesosphere/lower thermosphere wind time series." Universität Leipzig, 2010. https://ul.qucosa.de/id/qucosa%3A16361.
Full textAbstract:
A piecewise linear model is developed to detect climatic trends and possible structural changes in time series with a priori unknown number and positions of breakpoints. The initial noise is allowed to be interpreted by the first- and second-order autoregressive models. The goodness of fit of candidate models, if the residuals are accepted as normally distributed white noise, is evaluated using the Schwarz Bayesian Information Criterion. The uncertainties of all modeled trend parameters are estimated using the Monte-Carlo method. The model is applied to the mesosphere/lower thermosphere winds obtained at Collm (52°N, 15°E) during 1960-2007. A persistent increase after ~1980 is observed in the annual mean zonal wind based on the primary model while only a weak positive trend arises in the meridional component. Major trend breakpoints are identified around 1968-71 and 1976-79 in both the zonal and meridional winds.
29
Novacic, Jelena, and Kablai Tokhi. "Implementation of Anomaly Detection on a Time-series Temperature Data set." Thesis, Malmö universitet, Fakulteten för teknik och samhälle (TS), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:mau:diva-20375.
Full textAbstract:
Aldrig har det varit lika aktuellt med hållbar teknologi som idag. Behovet av bättre miljöpåverkan inom alla områden har snabbt ökat och energikonsumtionen är ett av dem. En enkel lösning för automatisk kontroll av energikonsumtionen i smarta hem är genom mjukvara. Med dagens IoT teknologi och maskinlärningsmodeller utvecklas den mjukvarubaserade hållbara livsstilen allt mer. För att kontrollera ett hushålls energikonsumption måste plötsligt avvikande beteenden detekteras och regleras för att undvika onödig konsumption. Detta examensarbete använder en tidsserie av temperaturdata för att implementera detektering av anomalier. Fyra modeller implementerades och testades; en linjär regressionsmodell, Pandas EWM funktion, en EWMA modell och en PEWMA modell. Varje modell testades genom att använda dataset från nio olika lägenheter, från samma tidsperiod. Därefter bedömdes varje modell med avseende på Precision, Recall och F-measure, men även en ytterligare bedömning gjordes för linjär regression med R^2-score. Resultaten visar att baserat på noggrannheten hos varje modell överträffade PEWMA de övriga modellerna. EWMA modeller var något bättre än den linjära regressionsmodellen, följt av Pandas egna EWM modell.Today's society has become more aware of its surroundings and the focus has shifted towards green technology. The need for better environmental impact in all areas is rapidly growing and energy consumption is one of them. A simple solution for automatically controlling the energy consumption of smart homes is through software. With today's IoT technology and machine learning models the movement towards software based ecoliving is growing. In order to control the energy consumption of a household, sudden abnormal behavior must be detected and adjusted to avoid unnecessary consumption. This thesis uses a time-series data set of temperature data for implementation of anomaly detection. Four models were implemented and tested; a Linear Regression model, Pandas EWM function, an exponentially weighted moving average (EWMA) model and finally a probabilistic exponentially weighted moving average (PEWMA) model. Each model was tested using data sets from nine different apartments, from the same time period. Then an evaluation of each model was conducted in terms of Precision, Recall and F-measure, as well as an additional evaluation for Linear Regression, using R^2 score. The results of this thesis show that in terms of accuracy, PEWMA outperformed the other models. The EWMA model was slightly better than the Linear Regression model, followed by the Pandas EWM model.
30
Aboode, Adam. "Anomaly Detection in Time Series Data Based on Holt-Winters Method." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-226344.
Full textAbstract:
In today's world the amount of collected data increases every day, this is a trend which is likely to continue. At the same time the potential value of the data does also increase due to the constant development and improvement of hardware and software. However, in order to gain insights, make decisions or train accurate machine learning models we want to ensure that the data we collect is of good quality. There are many definitions of data quality, in this thesis we focus on the accuracy aspect. One method which can be used to ensure accurate data is to monitor for and alert on anomalies. In this thesis we therefore suggest a method which, based on historic values, is able to detect anomalies in time series as new values arrive. The method consists of two parts, forecasting the next value in the time series using Holt-Winters method and comparing the residual to an estimated Gaussian distribution. The suggested method is evaluated in two steps. First, we evaluate the forecast accuracy for Holt-Winters method using different input sizes. In the second step we evaluate the performance of the anomaly detector when using different methods to estimate the variance of the distribution of the residuals. The results indicate that the suggested method works well most of the time for detection of point anomalies in seasonal and trending time series data. The thesis also discusses some potential next steps which are likely to further improve the performance of this method.I dagens värld ökar mängden insamlade data för varje dag som går, detta är en trend som sannolikt kommer att fortsätta. Samtidigt ökar även det potentiella värdet av denna data tack vare ständig utveckling och förbättring utav både hårdvara och mjukvara. För att utnyttja de stora mängder insamlade data till att skapa insikter, ta beslut eller träna noggranna maskininlärningsmodeller vill vi försäkra oss om att vår data är av god kvalité. Det finns många definitioner utav datakvalité, i denna rapport fokuserar vi på noggrannhetsaspekten. En metod som kan användas för att säkerställa att data är av god kvalité är att övervaka inkommande data och larma när anomalier påträffas. Vi föreslår därför i denna rapport en metod som, baserat på historiska data, kan detektera anomalier i tidsserier när nya värden anländer. Den föreslagna metoden består utav två delar, dels att förutsäga nästa värde i tidsserien genom Holt-Winters metod samt att jämföra residualen med en estimerad normalfördelning. Vi utvärderar den föreslagna metoden i två steg. Först utvärderas noggrannheten av de, utav Holt-Winters metod, förutsagda punkterna för olika storlekar på indata. I det andra steget utvärderas prestandan av anomalidetektorn när olika metoder för att estimera variansen av residualernas distribution används. Resultaten indikerar att den föreslagna metoden i de flesta fall fungerar bra för detektering utav punktanomalier i tidsserier med en trend- och säsongskomponent. I rapporten diskuteras även möjliga åtgärder vilka sannolikt skulle förbättra prestandan hos den föreslagna metoden.
31
Ferreira, Leonardo Nascimento. "Time series data mining using complex networks." Universidade de São Paulo, 2017. http://www.teses.usp.br/teses/disponiveis/55/55134/tde-01022018-144118/.
Full textAbstract:
A time series is a time-ordered dataset. Due to its ubiquity, time series analysis is interesting for many scientific fields. Time series data mining is a research area that is intended to extract information from these time-related data. To achieve it, different models are used to describe series and search for patterns. One approach for modeling temporal data is by using complex networks. In this case, temporal data are mapped to a topological space that allows data exploration using network techniques. In this thesis, we present solutions for time series data mining tasks using complex networks. The primary goal was to evaluate the benefits of using network theory to extract information from temporal data. We focused on three mining tasks. (1) In the clustering task, we represented every time series by a vertex and we connected vertices that represent similar time series. We used community detection algorithms to cluster similar series. Results show that this approach presents better results than traditional clustering results. (2) In the classification task, we mapped every labeled time series in a database to a visibility graph. We performed classification by transforming an unlabeled time series to a visibility graph and comparing it to the labeled graphs using a distance function. The new label is the most frequent label in the k-nearest graphs. (3) In the periodicity detection task, we first transform a time series into a visibility graph. Local maxima in a time series are usually mapped to highly connected vertices that link two communities. We used the community structure to propose a periodicity detection algorithm in time series. This method is robust to noisy data and does not require parameters. With the methods and results presented in this thesis, we conclude that network science is beneficial to time series data mining. Moreover, this approach can provide better results than traditional methods. It is a new form of extracting information from time series and can be easily extended to other tasks.Séries temporais são conjuntos de dados ordenados no tempo. Devido à ubiquidade desses dados, seu estudo é interessante para muitos campos da ciência. A mineração de dados temporais é uma área de pesquisa que tem como objetivo extrair informações desses dados relacionados no tempo. Para isso, modelos são usados para descrever as séries e buscar por padrões. Uma forma de modelar séries temporais é por meio de redes complexas. Nessa modelagem, um mapeamento é feito do espaço temporal para o espaço topológico, o que permite avaliar dados temporais usando técnicas de redes. Nesta tese, apresentamos soluções para tarefas de mineração de dados de séries temporais usando redes complexas. O objetivo principal foi avaliar os benefícios do uso da teoria de redes para extrair informações de dados temporais. Concentramo-nos em três tarefas de mineração. (1) Na tarefa de agrupamento, cada série temporal é representada por um vértice e as arestas são criadas entre as séries de acordo com sua similaridade. Os algoritmos de detecção de comunidades podem ser usados para agrupar séries semelhantes. Os resultados mostram que esta abordagem apresenta melhores resultados do que os resultados de agrupamento tradicional. (2) Na tarefa de classificação, cada série temporal rotulada em um banco de dados é mapeada para um gráfico de visibilidade. A classificação é realizada transformando uma série temporal não marcada em um gráfico de visibilidade e comparando-a com os gráficos rotulados usando uma função de distância. O novo rótulo é dado pelo rótulo mais frequente nos k grafos mais próximos. (3) Na tarefa de detecção de periodicidade, uma série temporal é primeiramente transformada em um gráfico de visibilidade. Máximos locais em uma série temporal geralmente são mapeados para vértices altamente conectados que ligam duas comunidades. O método proposto utiliza a estrutura de comunidades para realizar a detecção de períodos em séries temporais. Este método é robusto para dados ruidosos e não requer parâmetros. Com os métodos e resultados apresentados nesta tese, concluímos que a teoria da redes complexas é benéfica para a mineração de dados em séries temporais. Além disso, esta abordagem pode proporcionar melhores resultados do que os métodos tradicionais e é uma nova forma de extrair informações de séries temporais que pode ser facilmente estendida para outras tarefas.
32
Chen, Tiankai M. Eng Massachusetts Institute of Technology. "Anomaly detection in semiconductor manufacturing through time series forecasting using neural networks." Thesis, Massachusetts Institute of Technology, 2018. http://hdl.handle.net/1721.1/120245.
Full textAbstract:
Thesis: M. Eng. in Advanced Manufacturing and Design, Massachusetts Institute of Technology, Department of Mechanical Engineering, 2018.Cataloged from PDF version of thesis.
Includes bibliographical references (pages 92-94).
Semiconductor manufacturing provides unique challenges to the anomaly detection problem. With multiple recipes and multivariate data, it is difficult for engineers to reliably detect anomalies in the manufacturing process. An experimental study into anomaly detection through time series forecasting is carried out with application to a plasma etch case study. The study is performed on three predictive models with increasing complexity for comparison. The three models are namely: Autoregressive Integrated Moving Average (ARIMA), Multi-Layer Perceptron (MLP) and Long Short Term Memory (LSTM). ARIMA is a statistical model while MLP and LSTM are neural network models. The results from the control experiment, under supervised training, shows the validity of MLP and LSTM in detecting anomalies through time series forecasting with a recall accuracy of 92% for the best model. Conversely, the ARIMA model has a relatively poor performance due to the inability to model the data correctly. Experimental results also display the ability of neural network models to adapt to training sets of multiple recipes. Furthermore, downsampling is explored to reduce training times and has been found to have minor effects on the accuracy of the model. Moreover, an unsupervised approach towards anomaly detection is found to have little success in detecting anomalous points in the data.
by Tiankai Chen.
M. Eng. in Advanced Manufacturing and Design
33
Farahani, Marzieh. "Anomaly Detection on Gas Turbine Time-series’ Data Using Deep LSTM-Autoencoder." Thesis, Umeå universitet, Institutionen för datavetenskap, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-179863.
Full textAbstract:
Anomaly detection with the aim of identifying outliers plays a very important role in various applications (e.g., online spam, manufacturing, finance etc.). An automatic and reliable anomaly detection tool with accurate prediction is essential in many domains. This thesis proposes an anomaly detection method by applying deep LSTM (long short-term memory) especially on time-series data. By validating on real-worlddata at Siemens Industrial Turbomachinery (SIT), the proposed methods hows promising performance, and can be employed in different data domains like device logs of turbine machines to provide useful information on abnormal behaviors. In detail, our proposed method applies an auto encoder to have feature selection by keeping vital features, and learn the time series’s encoded representation. This approach reduces the extensive input data by pulling out the auto encoder’s latent layer output. For prediction, we then train a deep LSTM model with three hidden layers based on the encoder’s latent layer output. Afterwards, given the output from the prediction model, we detect the anomaly sensors related to the specific gas turbine by using a threshold approach. Our experimental results show that our proposed methods perform well on noisy and real-world data set in order to detect anomalies. Moreover, it confirmed that making predictions based on encoding representation, which is under reduction, is more accurate. We could say applying autoencoder can improve both anomaly detection and prediction tasks. Additionally, the performance of deep neural networks would be significantly improved for data with high complexity.
34
Chambon, Stanislas. "Learning from electrophysiology time series during sleep : from scoring to event detection." Thesis, Université Paris-Saclay (ComUE), 2018. http://www.theses.fr/2018SACLT014.
Full textAbstract:
Le sommeil est un phénomène biologique universel complexe et encore peu compris. La méthode de référence actuelle pour caractériser les états de vigilance au cours du sommeil est la polysomnographie (PSG) qui enregistre de manière non invasive à la surface de la peau, les modifications électrophysiologiques de l’activité cérébrale (électroencéphalographie, EEG), oculaire (électro-oculographie, EOG) et musculaire (électromyographie, EMG). Traditionnellement, les signaux électrophysiologiques sont ensuite analysés par un expert du sommeil qui annote manuellement les évènements d’intérêt comme les stades de sommeil ou certains micro-évènements (grapho éléments EEG). Toutefois, l’annotation manuelle est chronophage et sujette à la subjectivité de l’expert. De plus, le développement exponentiel d’outils de monitoring du sommeil enregistrant et analysant automatiquement les signaux électrophysiologiques tels que le bandeau Dreem rend nécessaire une automatisation de ces tâches.L’apprentissage machine bénéficie d’une attention croissante car il permet d’apprendre à un ordinateur à réaliser certaines tâches de décision à partir d’un ensemble d’exemples d’apprentissage et d’obtenir des performances de prédictions plus élevées qu’avec les méthodes classiques. Les avancées techniques dans le domaine de l’apprentissage profond ont ouvert de nouvelles perspectives pour la science du sommeil tout en soulevant de nouveaux défis techniques. L’entraînement des algorithmes d’apprentissage profond nécessite une grande quantité de données annotées qui n’est pas nécessairement disponible pour les données PSG. De plus, les algorithmes d’apprentissage sont très sensibles à la variabilité des données qui est non négligeable en ce qui concerne ces données. Cela s’explique par la variabilité intra et inter-sujet (pathologies / sujets sains, âge…).Cette thèse étudie le développement d’algorithmes d’apprentissage profond afin de réaliser deux types de tâches: la prédiction des stades de sommeil et la détection de micro-événements. Une attention particulière est portée (a) sur la quantité de données annotées requise pour l’entraînement des algorithmes proposés et (b) sur la sensibilité de ces algorithmes à la variabilité des données. Des stratégies spécifiques, basées sur l’apprentissage par transfert, sont proposées pour résoudre les problèmes techniques dus au manque de données annotées et à la variabilité des donnéesSleep is a complex and not fully understood biological phenomenon. The traditional process to monitor sleep relies on the polysomnography exam (PSG). It records, in a non invasive fashion at the level of the skin, electrophysiological modifications of the brain activity (electroencephalography, EEG), ocular (electro-oculography, EOG) and muscular (electro-myography, EMG). The recorded signals are then analyzed by a sleep expert who manually annotates the events of interest such as the sleep stages or some micro-events. However, manual labeling is time-consuming and prone to the expert subjectivity. Furthermore, the development of sleep monitoring consumer wearable devices which record and process automatically electrophysiological signals, such as Dreem headband, requires to automate some labeling tasks.Machine learning (ML) has received much attention as a way to teach a computer to perform some decision tasks automatically from a set of learning examples. Furthermore, the rise of deep learning (DL) algorithms in several fields have opened new perspectives for sleep sciences. On the other hand, this is also raising new concerns related to the scarcity of labeled data that may prevent their training processes and the variability of data that may hurt their performances. Indeed, sleep data is scarce due to the labeling burden and exhibits also some intra and inter-subject variability (due to sleep disorders, aging...).This thesis has investigated and proposed ML algorithms to automate the detection of sleep related events from raw PSG time series. Through the prism of DL, it addressed two main tasks: sleep stage classification and micro-event detection. A particular attention was brought (a) to the quantity of labeled data required to train such algorithms and (b) to the generalization performances of these algorithms to new (variable) data. Specific strategies, based on transfer learning, were designed to cope with the issues related to the scarcity of labeled data and the variability of data
35
Piyadi, Gamage Ramadha D. "Empirical Likelihood For Change Point Detection And Estimation In Time Series Models." Bowling Green State University / OhioLINK, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1495457528719879.
Full text
36
Jorge, Ana Maria Nabais. "Sobre a Definição de Outlier no Domínio Específico dos Modelos Lineares e Séries Temporais." Master's thesis, Instituto Superior de Economia e Gestão, 1999. http://hdl.handle.net/10400.5/4017.
Full textAbstract:
Mestrado em Matemática Aplicada à Economia e Gestão0 que são outliers e o que é o problema outlier? O principal objectivo desta dissertação é fazer o ponto da situação sobre o conceito outlier e como este é tratado no domínio específico dos modelos lineares e das séries temporais, para além de tentar saber como é que os mais modernos pacotes estatísticos tratam o tema. Assim, o nosso trabalho encontra-se dividido em seis capítulos. No primeiro capítulo, são enunciados os objectivos da presente dissertação e as suas motivações. No segundo capítulo, faz-se um apanhado de como o conceito outlier é definido pelos principais autores da especialidade assim como uma breve resenha histórica à evolução do tratamento dado a estas observações, desde Bernoulli até aos dias de hoje. O terceiro capítulo é dedicado à teoria geral de outliers, destacando-se os principais modelos e testes de discordância assim como alguns métodos de acomodação. O quarto capítulo é dedicado à análise do conceito outlier no âmbito dos modelos lineares. São referidos métodos de detecção, assim como alternativas robustas à estimação de mínimos quadrados. São ainda analisados vários pacotes estatísticos. O objectivo do quinto capítulo consiste no estudo de outliers em séries temporais. Serão classificados vários tipos de outliers. Será descrito e exem¬plificado um procedimento iterativo de detecção e estimação dos parâmetros da série proposto por Chang, Tiao e Chen (1988). Finalmente, no sexto capítulo extraem-se algumas conclusões sobre o tema.
What is an outlier and what's the outlier problem? The main purpose of this dissertation is to make the state of art about outlier's definition and how this subject is taking care in specific domain of linear models and time series besides trying to find out how statistical software treat the subject. So this dissertation is divided in six chapters. In the first chapter there is an explanation about the objectives of this dissertation, as well the motivations and importance of the subject. In the second chapter, there a review of how the major authors define the outlier concept as well as a brief historical reference to evolution of the treatment given to this observations since Bernoulli until now. The third chapter focus some aspects of the general theory of outliers, specially discordancy models and tests as well some accomodation proce¬dures. The fourth chapter regards the analysis of outlier's concept in the domain of linear models. It will be referenced some outlier's detection methods and some alternatives to least squares. They will be analysed several statistical packages. The objective of the fifth chapter is to study outliers in time series. Several types of outliers will be classified. It will be explain and illustrad a iterative procedure for outlier detection and parameter estimation develop by Chang, Tiao and Chen (1988). Finally, in the sixth chapter, some conclusions are draw.
37
Ravirala, Narayana. "Device signal detection methods and time frequency analysis." Diss., Rolla, Mo. : University of Missouri-Rolla, 2007. http://scholarsmine.umr.edu/thesis/pdf/Ravirala_09007dcc803fea67.pdf.
Full textAbstract:
Thesis (M.S.)--University of Missouri--Rolla, 2007.Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed March 18, 2008) Includes bibliographical references (p. 89-90).
38
Alklid, Jonathan. "Time to Strike: Intelligent Detection of Receptive Clients : Predicting a Contractual Expiration using Time Series Forecasting." Thesis, Linnéuniversitetet, Institutionen för datavetenskap och medieteknik (DM), 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-106217.
Full textAbstract:
In recent years with the advances in Machine Learning and Artificial Intelligence, the demand for ever smarter automation solutions could seem insatiable. One such demand was identified by Fortnox AB, but undoubtedly shared by many other industries dealing with contractual services, who were looking for an intelligent solution capable of predicting the expiration date of a contractual period. As there was no clear evidence suggesting that Machine Learning models were capable of learning the patterns necessary to predict a contract's expiration, it was deemed desirable to determine subject feasibility while also investigating whether it would perform better than a commonplace rule-based solution, something that Fortnox had already investigated in the past. To do this, two different solutions capable of predicting a contractual expiration were implemented. The first one was a rule-based solution that was used as a measuring device, and the second was a Machine Learning-based solution that featured Tree Decision classifier as well as Neural Network models. The results suggest that Machine Learning models are indeed capable of learning and recognizing patterns relevant to the problem, and with an average accuracy generally being on the high end. Unfortunately, due to a lack of available data to use for testing and training, the results were too inconclusive to make a reliable assessment of overall accuracy beyond the learning capability. The conclusion of the study is that Machine Learning-based solutions show promising results, but with the caveat that the results should likely be seen as indicative of overall viability rather than representative of actual performance.
39
Theissler, Andreas. "Detecting anomalies in multivariate time series from automotive systems." Thesis, Brunel University, 2013. http://bura.brunel.ac.uk/handle/2438/7902.
Full textAbstract:
In the automotive industry test drives are conducted during the development of new vehicle models or as a part of quality assurance for series vehicles. During the test drives, data is recorded for the use of fault analysis resulting in millions of data points. Since multiple vehicles are tested in parallel, the amount of data that is to be analysed is tremendous. Hence, manually analysing each recording is not feasible. Furthermore the complexity of vehicles is ever-increasing leading to an increase of the data volume and complexity of the recordings. Only by effective means of analysing the recordings, one can make sure that the effort put in the conducting of test drives pays off. Consequently, effective means of test drive analysis can become a competitive advantage. This Thesis researches ways to detect unknown or unmodelled faults in recordings from test drives with the following two aims: (1) in a data base of recordings, the expert shall be pointed to potential errors by reporting anomalies, and (2) the time required for the manual analysis of one recording shall be shortened. The idea to achieve the first aim is to learn the normal behaviour from a training set of recordings and then to autonomously detect anomalies. The one-class classifier “support vector data description” (SVDD) is identified to be most suitable, though it suffers from the need to specify parameters beforehand. One main contribution of this Thesis is a new autonomous parameter tuning approach, making SVDD applicable to the problem at hand. Another vital contribution is a novel approach enhancing SVDD to work with multivariate time series. The outcome is the classifier “SVDDsubseq” that is directly applicable to test drive data, without the need for expert knowledge to configure or tune the classifier. The second aim is achieved by adapting visual data mining techniques to make the manual analysis of test drives more efficient. The methods of “parallel coordinates” and “scatter plot matrices” are enhanced by sophisticated filter and query operations, combined with a query tool that allows to graphically formulate search patterns. As a combination of the autonomous classifier “SVDDsubseq” and user-driven visual data mining techniques, a novel, data-driven, semi-autonomous approach to detect unmodelled faults in recordings from test drives is proposed and successfully validated on recordings from test drives. The methodologies in this Thesis can be used as a guideline when setting up an anomaly detection system for own vehicle data.
40
Eriksson, Tilda. "Change Detection in Telecommunication Data using Time Series Analysis and Statistical Hypothesis Testing." Thesis, Linköpings universitet, Matematiska institutionen, 2013. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-94530.
Full textAbstract:
In the base station system of the GSM mobile network there are a large number of counters tracking the behaviour of the system. When the software of the system is updated, we wish to find out which of the counters that have changed their behaviour. This thesis work has shown that the counter data can be modelled as a stochastic time series with a daily profile and a noise term. The change detection can be done by estimating the daily profile and the variance of the noise term and perform statistical hypothesis tests of whether the mean value and/or the daily profile of the counter data before and after the software update can be considered equal. When the chosen counter data has been analysed, it seems to be reasonable in most cases to assume that the noise terms are approximately independent and normally distributed, which justies the hypothesis tests. When the change detection is tested on data where the software is unchanged and on data with known software updates, the results are as expected in most cases. Thus the method seems to be applicable under the conditions studied.
41
Kwong, Siu-shing. "Detection of determinism of nonlinear time series with application to epileptic electroencephalogram analysis." View the Table of Contents & Abstract, 2005. http://sunzi.lib.hku.hk/hkuto/record/B35512222.
Full text
42
Oliveira, Adriano Lorena Inácio de. "Neural networks forecasting and classification-based techniques for novelty detection in time series." Universidade Federal de Pernambuco, 2011. https://repositorio.ufpe.br/handle/123456789/1825.
Full textAbstract:
Made available in DSpace on 2014-06-12T15:52:37Z (GMT). No. of bitstreams: 2
arquivo4525_1.pdf: 1657788 bytes, checksum: 5abba3555b6cbbc4fa073f1b718d6579 (MD5)
license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5)
Previous issue date: 2011O problema da detecção de novidades pode ser definido como a identificação de dados novos ou desconhecidos aos quais um sistema de aprendizagem de máquina não teve acesso durante o treinamento. Os algoritmos para detecção de novidades são projetados para classificar um dado padrão de entrada como normal ou novidade. Esses algoritmos são usados em diversas areas, como visão computacional, detecçãao de falhas em máquinas, segurança de redes de computadores e detecção de fraudes. Um grande número de sistemas pode ter seu comportamento modelado por séries temporais. Recentemente o pro oblema de detecção de novidades em séries temporais tem recebido considerável atenção. Várias técnicas foram propostas, incluindo téecnicas baseadas em previsão de séries temporais com redes neurais artificiais e em classificação de janelas das s´eries temporais. As t´ecnicas de detec¸c ao de novidades em s´eries temporais atrav´es de previs ao t em sido criticadas devido a seu desempenho considerado insatisfat´orio. Em muitos problemas pr´aticos, a quantidade de dados dispon´ıveis nas s´eries ´e bastante pequena tornando a previs ao um problema ainda mais complexo. Este ´e o caso de alguns problemas importantes de auditoria, como auditoria cont´abil e auditoria de folhas de pagamento. Como alternativa aos m´etodos baseados em previs ao, alguns m´etodos baseados em classificação foram recentemente propostos para detecção de novidades em séries temporais, incluindo m´etodos baseados em sistemas imunol´ogicos artificiais, wavelets e m´aquinas de vetor de suporte com uma ´unica classe. Esta tese prop oe um conjunto de m´etodos baseados em redes neurais artificiais para detecção de novidades em séries temporais. Os métodos propostos foram projetados especificamente para detec¸c ao de fraudes decorrentes de desvios relativamente pequenos, que s ao bastante importantes em aplica¸c oes de detec¸c ao de fraudes em sistemas financeiros. O primeiro m´etodo foi proposto para melhorar o desempenho de detec¸c ao de novidades baseada em previs ao. Este m´etodo ´e baseado em intervalos de confian¸ca robustos, que s ao usados para definir valores adequados para os limiares a serem usados para detec¸c ao de novidades. O m´etodo proposto foi aplicado a diversas s´eries temporais financeiras e obteve resultados bem melhores que m´etodos anteriores baseados em previs ao. Esta tese tamb´em prop oe dois diferentes m´etodos baseados em classifica¸c ao para detec ¸c ao de novidades em s´eries temporais. O primeiro m´etodo ´e baseado em amostras negativas, enquanto que o segundo m´etodo ´e baseado em redes neurais artificiais RBFDDA e n ao usa amostras negativas na fase de treinamento. Resultados de simula¸c ao usando diversas s´eries temporais extra´ıdas de aplica¸c oes reais mostraram que o segundo m´etodo obt´em melhor desempenho que o primeiro. Al´em disso, o desempenho do segundo m´etodo n ao depende do tamanho do conjunto de teste, ao contr´ario do que acontece com o primeiro m´etodo. Al´em dos m´etodos para detec¸c ao de novidades em s´eries temporais, esta tese prop oe e investiga quatro diferentes m´etodos para melhorar o desempenho de redes neurais RBF-DDA. Os m´etodos propostos foram avaliados usando seis conjuntos de dados do reposit´orio UCI e os resultados mostraram que eles melhoram consideravelmente o desempenho de redes RBF-DDA e tamb´em que eles obt em melhor desempenho que redes MLP e que o m´etodo AdaBoost. Al´em disso, mostramos que os m´etodos propostos obt em resultados similares a k-NN. Os m´etodos propostos para melhorar RBF-DDA foram tamb´em usados em conjunto com o m´etodo proposto nesta tese para detec¸c ao de novidades em s´eries temporais baseado em amostras negativas. Os resultados de diversos experimentos mostraram que esses m´etodos tamb´em melhoram bastante o desempenho da detec¸c ao de fraudes em s´eries temporais, que ´e o foco principal desta tese.
43
Grobler, Trienko Lups. "Sequential and non-sequential hypertemporal classification and change detection of Modis time-series." Thesis, University of Pretoria, 2012. http://hdl.handle.net/2263/25427.
Full textAbstract:
Satellites provide humanity with data to infer properties of the earth that were impossible a century ago. Humanity can now easily monitor the amount of ice found on the polar caps, the size of forests and deserts, the earth’s atmosphere, the seasonal variation on land and in the oceans and the surface temperature of the earth. In this thesis, new hypertemporal techniques are proposed for the settlement detection problem in South Africa. The hypertemporal techniques are applied to study areas in the Gauteng and Limpopo provinces of South Africa. To be more specific, new sequential (windowless) and non-sequential hypertemporal techniques are implemented. The time-series employed by the new hypertemporal techniques are obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor, which is on board the earth observations satellites Aqua and Terra. One MODIS dataset is constructed for each province. A Support Vector Machine (SVM) [1] that uses a novel noise-harmonic feature set is implemented to detect existing human settlements. The noise-harmonic feature set is a non-sequential hypertemporal feature set and is constructed by using the Coloured Simple Harmonic Oscillator (CSHO) [2]. The CSHO consists of a Simple Harmonic Oscillator (SHO) [3], which is superimposed on the Ornstein-Uhlenbeck process [4]. The noise-harmonic feature set is an extension of the classic harmonic feature set [5]. The classic harmonic feature set consists of a mean and a seasonal component. For the case studies in this thesis, it is observed that the noise-harmonic feature set not only extends the harmonic feature set, but also improves on its classification capability. The Cumulative Sum (CUSUM) algorithm was developed by Page in 1954 [6]. In its original form it is a sequential (windowless) hypertemporal change detection technique. Windowed versions of the algorithm have been applied in a remote sensing context. In this thesis CUSUM is used in its original form to detect settlement expansion in South Africa and is benchmarked against the classic band differencing change detection approach of Lunetta et al., which was developed in 2006 [7]. In the case of the Gauteng study area, the CUSUM algorithm outperformed the band differencing technique. The exact opposite behaviour was seen in the case of the Limpopo dataset. Sequential hypertemporal techniques are data-intensive and an inductive MODIS simulator was therefore also developed (to augment datasets). The proposed simulator is also based on the CSHO. Two case studies showed that the proposed inductive simulator accurately replicates the temporal dynamics and spectral dependencies found in MODIS data.Thesis (PhD(Eng))--University of Pretoria, 2012.
Electrical, Electronic and Computer Engineering
unrestricted
44
Biswas, Debashis. "An Algorithm for Mining Adverse-Event Datasets for Detection of Post Safety Concern of a Drug." Scholarly Repository, 2010. http://scholarlyrepository.miami.edu/oa_theses/17.
Full textAbstract:
Signal detection from Adverse Event Reports (AERs) is important for identifying and analysing drug safety concern after a drug has been released into the market. A safety signal is defined as a possible causal relation between an adverse event and a drug. There are a number of safety signal detection algorithms available for detecting drug safety concern. They compare the ratio of observed count to expected count to find instances of disproportionate reportings of an event for a drug or combination of events for a drug. In this thesis, we present an algorithm to mine the AERs to identify drugs which show sudden and large changes in patterns of reporting of adverse events. Unlike other algorithms, the proposed algorithm creates time series for each drug and use it to identify start of a potential safety problem. A novel vectorized timeseries utilizing multiple attributes has been proposed here. First a time series with a small time period was created; then to remove local variations of the number of reports in a time period, a time-window based averaging was done. This method helped to keep a relatively long time-series, but eliminated local variations. The steps in the algorithm include partitioning the counts on attribute values, creating a vector out of the partitioned counts for each time period, use of a sliding time window, normalizing the vectors and computing vector differences to find the changes in reporting over time. Weights have been assigned to attributes to highlight changes in the more significant attributes. The algorithm was tested with Adverse Event Reporting System (AERS) datasets from Food and Drug Administation (FDA). From AERS datasets the proposed algorithm identified five drugs that may have safety concern. After searching literature and the Internet it was found that the five drugs the algorithm identified, two were recalled, one was suspended, one had to undergo label change and the other one has a lawsuit pending against it.
45
Mohr, Maria [Verfasser], and Natalie [Akademischer Betreuer] Neumeyer. "Changepoint detection in a nonparametric time series regression model / Maria Mohr ; Betreuer: Natalie Neumeyer." Hamburg : Staats- und Universitätsbibliothek Hamburg, 2018. http://d-nb.info/1171988303/34.
Full text
46
Bracci, Lorenzo, and Amirhossein Namazi. "EVALUATION OF UNSUPERVISED MACHINE LEARNING MODELS FOR ANOMALY DETECTION IN TIME SERIES SENSOR DATA." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-299734.
Full textAbstract:
With the advancement of the internet of things and the digitization of societies sensor recording time series data can be found in an always increasing number of places including among other proximity sensors on cars, temperature sensors in manufacturing plants and motion sensors inside smart homes. This always increasing reliability of society on these devices lead to a need for detecting unusual behaviour which could be caused by malfunctioning of the sensor or by the detection of an uncommon event. The unusual behaviour mentioned is often referred to as an anomaly. In order to detect anomalous behaviours, advanced technologies combining mathematics and computer science, which are often referred to as under the umbrella of machine learning, are frequently used to solve these problems. In order to help machines to learn valuable patterns often human supervision is needed, which in this case would correspond to use recordings which a person has already classified as anomalies or normal points. It is unfortunately time consuming to label data, especially the large datasets that are created from sensor recordings. Therefore in this thesis techniques that require no supervision are evaluated to perform anomaly detection. Several different machine learning models are trained on different datasets in order to gain a better understanding concerning which techniques perform better when different requirements are important such as presence of a smaller dataset or stricter requirements on inference time. Out of the models evaluated, OCSVM resulted in the best overall performance, achieving an accuracy of 85% and K- means was the fastest model as it took 0.04 milliseconds to run inference on one sample. Furthermore LSTM based models showed most possible improvements with larger datasets.Med utvecklingen av Sakernas internet och digitaliseringen av samhället kan man registrera tidsseriedata på allt fler platser, bland annat igenom närhetssensorer på bilar, temperatursensorer i tillverkningsanläggningar och rörelsesensorer i smarta hem. Detta ständigt ökande beroende i samhället av dessa enheter leder till ett behov av att upptäcka ovanligt beteende som kan orsakas av funktionsstörning i sensorn eller genom upptäckt av en ovanlig händelse. Det ovanliga beteendet som nämns kallas ofta för en anomali. För att upptäcka avvikande beteenden används avancerad teknik som kombinerar matematik och datavetenskap, som ofta kallas maskininlärning. För att hjälpa maskiner att lära sig värdefulla mönster behövs ofta mänsklig tillsyn, vilket i detta fall skulle motsvara användningsinspelningar som en person redan har klassificerat som avvikelser eller normala punkter. Tyvärr är det tidskrävande att märka data, särskilt de stora datamängder som skapas från sensorinspelningar. Därför utvärderas tekniker som inte kräver någon handledning i denna avhandling för att utföra anomalidetektering. Flera olika maskininlärningsmodeller utbildas på olika datamängder för att få en bättre förståelse för vilka tekniker som fungerar bättre när olika krav är viktiga, t.ex. närvaro av en mindre dataset eller strängare krav på inferens tid. Av de utvärderade modellerna resulterade OCSVM i bästa totala prestanda, uppnådde en noggrannhet på 85% och K- means var den snabbaste modellen eftersom det hade en inferens tid av 0,04 millisekunder. Dessutom visade LSTM- baserade modeller de bästa möjliga förbättringarna med större datamängder.
47
Tian, Runfeng. "An Enhanced Approach using Time Series Segmentation for Fault Detection of Semiconductor Manufacturing Process." University of Cincinnati / OhioLINK, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1562923441016763.
Full text
48
Dou, Baojun. "Three essays on time series : spatio-temporal modelling, dimension reduction and change-point detection." Thesis, London School of Economics and Political Science (University of London), 2015. http://etheses.lse.ac.uk/3242/.
Full textAbstract:
Modelling high dimensional time series and non-stationary time series are two import aspects in time series analysis nowadays. The main objective of this thesis is to deal with these two problems. The first two parts deal with high dimensionality and the third part considers a change point detection problem. In the first part, we consider a class of spatio-temporal models which extend popular econometric spatial autoregressive panel data models by allowing the scalar coefficients for each location (or panel) different from each other. The model is of the following form: yt = D(λ0)Wyt + D(λ1)yt−1 + D(λ2)Wyt−1 + εt, (1) where yt = (y1,t, . . . , yp,t) T represents the observations from p locations at time t, D(λk) = diag(λk1, . . . , λkp) and λkj is the unknown coefficient parameter for the j-th location, and W is the p×p spatial weight matrix which measures the dependence among different locations. All the elements on the main diagonal of W are zero. It is a common practice in spatial econometrics to assume W known. For example, we may let wij = 1/(1 + dij ), for i ̸= j, where dij ≥ 0 is an appropriate distance between the i-th and the j-th location. It can simply be the geographical distance between the two locations or the distance reflecting the correlation or association between the variables at the two locations. In the above model, D(λ0) captures the pure spatial effect, D(λ1) captures the pure dynamic effect, and D(λ2) captures the time-lagged spatial effect. We also assume that the error term εt = (ε1,t, ε2,t, . . . , εp,t) T in (1) satisfies the condition Cov (yt−1, εt) = 0. When λk1 = · · · = λkp for all k = 1, 2, 3, (1) reduces to the model of Yu et al. (2008), in which there are only 3 unknown regressive coefficient parameters. In general the regression function in (1) contains 3p unknown parameters. To overcome the innate endogeneity, we propose a generalized Yule-Walker estimation method which applies the least squares estimation to a Yule-Walker equation. The asymptotic theory is developed under the setting that both the sample size and the number of locations (or panels) tend to infinity under a general setting for stationary and α-mixing processes, which includes spatial autoregressive panel data models driven by i.i.d. innovations as special cases. The proposed methods are illustrated using both simulated and real data. In part 2, we consider a multivariate time series model which decomposes a vector process into a latent factor process and a white noise process. Let yt = (y1,t, · · · , yp,t) T be an observable p × 1 vector time series process. The factor model decomposes yt in the following form: yt = Axt + εt , (2) where xt = (x1,t, · · · , xr,t) T is a r × 1 latent factor time series with unknown r ≤ p and A = (a1, a2, · · · , ar) is a p × r unknown constant matrix. εt is a white noise process with mean 0 and covariance matrix Σε. The first part of (2) is a dynamic part and the serial dependence of yt is driven by xt. We will achieve dimension reduction once r ≪ p in the sense that the dynamics of yt is driven by a much lower dimensional process xt. Motivated by practical needs and the characteristic of high dimensional data, the sparsity assumption on factor loading matrix is imposed. Different from Lam, Yao and Bathia (2011)’s method, which is equivalent to an eigenanalysis of a non negative definite matrix, we add a constraint to control the number of nonzero elements in each column of the factor loading matrix. Our proposed sparse estimator is then the solution of a constrained optimization problem. The asymptotic theory is developed under the setting that both the sample size and the dimensionality tend to infinity. When the common factor is weak in the sense that δ > 1/2 in Lam, Yao and Bathia (2011)’s paper, the new sparse estimator may have a faster convergence rate. Numerically, we employ the generalized deflation method (Mackey (2009)) and the GSLDA method (Moghaddam et al. (2006)) to approximate the estimator. The tuning parameter is chosen by cross validation. The proposed method is illustrated with both simulated and real data examples. The third part is a change point detection problem. we consider the following covariance structural break detection problem: Cov(yt)I(tj−1 ≤ t < tj ) = Σtj−1, j = 1, · · · , m + 1, where yt is a p × 1 vector time series, Σtj−1̸ = Σtj and {t1, . . ., tm} are change points, 1 = t0 < t1 < · · · < tm+1 = n. In the literature, the number of change points m is usually assumed to be known and small, because a large m would involve a huge amount of computational burden for parameters estimation. By reformulating the problem in a variable selection context, the group least absolute shrinkage and selection operator (LASSO) is proposed to estimate m and the locations of the change points {t1, . . ., tm}. Our method is model free, it can be extensively applied to multivariate time series, such as GARCH and stochastic volatility models. It is shown that both m and the locations of the change points {t1, . . . , tm} can be consistently estimated from the data, and the computation can be efficiently performed. An improved practical version that incorporates group LASSO and the stepwise regression variable selection technique are discussed. Simulation studies are conducted to assess the finite sample performance.
49
Mohr, Maria Verfasser], and Natalie [Akademischer Betreuer] [Neumeyer. "Changepoint detection in a nonparametric time series regression model / Maria Mohr ; Betreuer: Natalie Neumeyer." Hamburg : Staats- und Universitätsbibliothek Hamburg, 2018. http://nbn-resolving.de/urn:nbn:de:gbv:18-94167.
Full text
50
Kayastha, Nilam. "Application on Lidar and Time Series Landsat Data for Mapping and Monitoring Wetlands." Diss., Virginia Tech, 2014. http://hdl.handle.net/10919/54011.
Full textAbstract:
To successfully protect and manage wetlands, efficient and accurate tools are needed to identify where wetlands are located, the wetland type, what condition they are in, what are the stressors present, and the trend in their condition. Wetland mapping and monitoring are useful to accomplish these tasks. Wetland mapping and monitoring with optical remote sensing data has mainly focused on using a single image or using image acquired over two seasons within the same year. Now that Landsat data are available freely, a multi-temporal approach utilizing images that span multiple seasons and multiple years can potentially be used to characterize wetland dynamics in more detail. In addition, newer remote sensing techniques such as Light Detection and Ranging (lidar) can provide highly detailed and accurate topographic information, which can improve our ability to discriminate wetlands. Thus, the overall objective of this study was to investigate the utility of lidar and multi-temporal Landsat data for mapping and monitoring of wetlands. My research is presented as three independent studies related to wetland mapping and monitoring. In the first study, inter-annual time series of Landsat data from 1985 to 2009 was used to map changes in wetland ecosystems in northern Virginia. Z-scores calculated on tasseled cap images were used to develop temporal profile for wetlands delineated by the National Wetland Inventory. A change threshold was derived based on the Chi-square distribution of the Z-scores. The accuracy of a change/no change map produced was 89% with a
kappa value of 0.79. Assessment of the change map showed that the method used was able to detect complete wetland loss together with other subtle changes resulting from development, harvesting, thinning and farming practices. The objective of the second study was to characterize differences in spectro-temporal profile of forested upland and wetland using intra and inter annual time series of Landsat data (1999-2012). The results show that the spector-temporal metrics derived from Landsat can accurately discriminate between forested upland and wetland (accuracy of 88.5%). The objective of the third study was to investigate the ability of topographic variables derived from lidar to map wetlands. Different topographic variables were derived from a high resolution lidar digital elevation model. Random forest model was used to assess the ability of these variables in mapping wetlands and uplands area. The result shows that lidar data can discriminate between wetlands and uplands with an accuracy of 72%. In summary, because of its spatial, spectral, temporal resolution, availability and cost Landsat data will be a primary data source for mapping and monitoring wetlands. The multi-temporal approach presented in this study has great potential for significantly improving our ability to detect and monitor wetlands. In addition, synergistic use of multi-temporal analysis of Landsat data combined with lidar data may be superior to using either data alone for future wetland mapping and monitoring approaches.Ph. D.