Dissertations / Theses on the topic 'ARIMA'
To see the other types of publications on this topic, follow the link: ARIMA.
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 'ARIMA.'
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
Abalos, Choque Melisa. "Modelo Arima con intervenciones." Universidad Mayor de San Andrés. Programa Cybertesis BOLIVIA, 2009. http://www.cybertesis.umsa.bo:8080/umsa/2009/abalos_cme/html/index-frames.html.
Full textAbstract:
El desarrollo de gran parte de los modelos y métodos estadísticos, específicamente relacionados con series temporales, ha ido ligado al deseo de estudiar aplicaciones específicas dentro de diversos ámbitos científicos. El presente trabajo también surgió con el objetivo de resolver diversos problemas que se plantean dentro del ámbito econométrico, aunque también puede ser usado en otros ámbitos, todos ellos ligados con un conjunto de datos históricos y con una aplicación muy concreta al estudio del “egreso de divisas” en Bolivia. Se han estudiado a profundidad los modelos para series temporales que únicamente dependían del pasado de la propia serie. En el presente trabajo se inicia el análisis de una serie temporal teniendo en cuenta algún tipo de información externa. En el capítulo 1 se sustenta fuertemente el hecho de investigar acerca de aspectos ajenos a la serie temporal que llegan de algún modo a alterar su normal comportamiento. El capítulo 2 desarrolla minuciosamente modelos univariantes conocidos con el nombre de ARIMA, desarrollando su parte teórica. Posteriormente se complementa esta perspectiva univariante añadiéndose una parte determinística correspondiente al análisis de intervención construyendo así el modelo ARIMA CON INTERVENCIONES, la utilización de éstos modelos es comparada en el capítulo 3, de esta manera se distingui cual de los dos es más efectivo cuando los datos son afectados por eventos circunstanciales. La metodología del modelo ARIMA CON INTERVENCIONES es una herramienta útil para “modelizar” el comportamiento de las series temporales que presentan modificaciones a raíz de eventos ajenos que no pueden ser controlados.
2
Rostami, Tabar Bahman. "ARIMA demand forecasting by aggregation." Phd thesis, Université Sciences et Technologies - Bordeaux I, 2013. http://tel.archives-ouvertes.fr/tel-00980614.
Full textAbstract:
Demand forecasting performance is subject to the uncertainty underlying the time series an organisation is dealing with. There are many approaches that may be used to reduce demand uncertainty and consequently improve the forecasting (and inventory control) performance. An intuitively appealing such approach that is known to be effective is demand aggregation. One approach is to aggregate demand in lower-frequency 'time buckets'. Such an approach is often referred to, in the academic literature, as temporal aggregation. Another approach discussed in the literature is that associated with cross-sectional aggregation, which involves aggregating different time series to obtain higher level forecasts.This research discusses whether it is appropriate to use the original (not aggregated) data to generate a forecast or one should rather aggregate data first and then generate a forecast. This Ph.D. thesis reveals the conditions under which each approach leads to a superior performance as judged based on forecast accuracy. Throughout this work, it is assumed that the underlying structure of the demand time series follows an AutoRegressive Integrated Moving Average (ARIMA) process.In the first part of our1 research, the effect of temporal aggregation on demand forecasting is analysed. It is assumed that the non-aggregate demand follows an autoregressive moving average process of order one, ARMA(1,1). Additionally, the associated special cases of a first-order autoregressive process, AR(1) and a moving average process of order one, MA(1) are also considered, and a Single Exponential Smoothing (SES) procedure is used to forecast demand. These demand processes are often encountered in practice and SES is one of the standard estimators used in industry. Theoretical Mean Squared Error expressions are derived for the aggregate and the non-aggregate demand in order to contrast the relevant forecasting performances. The theoretical analysis is validated by an extensive numerical investigation and experimentation with an empirical dataset. The results indicate that performance improvements achieved through the aggregation approach are a function of the aggregation level, the smoothing constant value used for SES and the process parameters.In the second part of our research, the effect of cross-sectional aggregation on demand forecasting is evaluated. More specifically, the relative effectiveness of top-down (TD) and bottom-up (BU) approaches are compared for forecasting the aggregate and sub-aggregate demands. It is assumed that that the sub-aggregate demand follows either a ARMA(1,1) or a non-stationary Integrated Moving Average process of order one, IMA(1,1) and a SES procedure is used to extrapolate future requirements. Such demand processes are often encountered in practice and, as discussed above, SES is one of the standard estimators used in industry (in addition to being the optimal estimator for an IMA(1) process). Theoretical Mean Squared Errors are derived for the BU and TD approach in order to contrast the relevant forecasting performances. The theoretical analysis is supported by an extensive numerical investigation at both the aggregate and sub-aggregate levels in addition to empirically validating our findings on a real dataset from a European superstore. The results show that the superiority of each approach is a function of the series autocorrelation, the cross-correlation between series and the comparison level.Finally, for both parts of the research, valuable insights are offered to practitioners and an agenda for further research in this area is provided.
3
Mariotti, Mara Terezinha. "Análise arima de dados meteo-oceanográficos." Florianópolis, SC, 2003. http://repositorio.ufsc.br/xmlui/handle/123456789/84655.
Full textAbstract:
Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico. Programa de Pós-Graduação em Engenharia Ambiental.Made available in DSpace on 2012-10-20T11:18:26Z (GMT). No. of bitstreams: 1 196113.pdf: 2449450 bytes, checksum: 56b01ccb27ccad38aa728c00dc2045ee (MD5)
Um estudo do mecanismo gerador das componentes meteorológicas que afetam o nível do mar é proposto através da utilização de modelos ARIMA (autorregressive integrated moving average). Séries temporais da temperatura do ar, pressão atmosférica, da componente meridional do vento e do nível do mar foram aquisitadas em São Francisco do Sul-SC, no período de 14 de julho a 15 de dezembro de 1996, e reamostradas a cada seis horas para melhor avaliar as componentes de baixa freqüência. As séries se mostraram não estacionárias na média, impondo a necessidade de integração. Não foi possível identificar uma não estacionaridade da variância devido ao comprimento insuficiente dos registros utilizados. Nos modelos de ordem 2 a estrutura de recorrência entre dois sistemas frontais é reconhecida através do modo associado aos dois pólos do polinômio. Os modelos AR(4) de todas as variáveis consideradas conseguem reconstruir também a evolução do sistema in situ, de período aproximado de 2,5 dias, por meio da segunda dupla de pólos. Modelos autorregressivos de ordem superior poderiam melhorar a identificação e a reconstrução desses ciclos, mas não conseguem convergir devido a não estacionaridade. Apesar disso, modelos de baixa ordem, com dois parâmetros apenas, conseguem fazer previsões aceitáveis até 24 horas, o que demonstra as possibilidades da metodologia.
4
Örneholm, Filip. "Anomaly Detection in Seasonal ARIMA Models." Thesis, Uppsala universitet, Tillämpad matematik och statistik, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-388503.
Full text
5
Vollenbröker, Bernd Karl [Verfasser], and Alexander [Akademischer Betreuer] Lindner. "Strictly Stationary Solutions of Multivariate ARMA and Univariate ARIMA Equations / Bernd Karl Vollenbröker ; Betreuer: Alexander Lindner." Braunschweig : Technische Universität Braunschweig, 2011. http://d-nb.info/1175824860/34.
Full text
6
Guimarães, Rita Cabral Pereira de Castro. "Modelização ARIMA de sucessões cronológicas: aplicação na previsão de escoamentos mensais." Master's thesis, Universidade de Évora, 1997. http://hdl.handle.net/10174/13282.
Full textAbstract:
Os modelos ARIMA tem vindo a ser cada vez mais utilizados na modelização e previsão de sucessões hidrológicas, instrumento fundamental para o planeamento e gestão de qualquer sistema do domínio Hídrico. A modelização de tais sucessões é conseguida através de uma metodologia em três etapas, desenvolvida por G. E. P. Box e G. M. Jenkins. Deste processo resulta um modelo, considerado como o mais adequado para representar a sucessão, podendo este ser então utilizado na previsão de eventos futuros. Para a aplicação destes modelos utilizaram-se seis sucessões de escoamentos mensais observados em três cursos de água pertencentes à bacia hidrográfica do Rio Douro. A modelização efectuada para esta sucessões permitiu eleger, para cada uma delas, um modelo ARIMA, com o qual se estabeleceram previsões para dois anos consecutivos à última observação. / Abstract - ARIMA models have become an important tool for modelling and forecasting of hydrologic sequences. Theses techniques are of considerable importance to the design and operation of water resource systems. Before being able to forecasting future values, models have to be found which describe past data adequately. These is accomplished with a iterative process, developed by G. E. P. Box and G. M. Jenkins, which incorporates three stages. For the applications of these models we selected six monthly flow sequences for three rivers located in Douro River watershed. The modelling of such sequences gave one ARIMA model for the forecasting of flows two years ahead.
7
Філатова, Ганна Петрівна, Анна Петровна Филатова, and Hanna Petrivna Filatova. "Прогнозування державного боргу з використанням ARIMA моделі." Thesis, ЦФЕНД, 2020. https://essuir.sumdu.edu.ua/handle/123456789/84293.
Full textAbstract:
Державний борг як важливий фактор соціально-економічного розвитку держави виступає свого роду індикатором і критерієм ефективності провадження виваженої боргової політики держави, а його прогнозування займає одне з ключових місць в процесі забезпечення економічної безпеки держави. У сучасній статистичній теорії існує безліч різноманітних методів прогнозування економічної інформації. Значна їх частина стосується прогнозування часових рядів, без додаткової інформації, тобто без аналізу впливу інших факторів. Звичайно, такий аналіз є доволі неповним, але досить часто результати таких прогнозів є більш точними порівняно з іншими методами прогнозування. Одним з таких методів є побудова ARIMA моделі.
8
Muller, Daniela. "Estimação para os parâmetros de processos estocásticos estacionários com característica de longa dependência." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 1999. http://hdl.handle.net/10183/127017.
Full textAbstract:
Estudos recentes em séries temporais direcionam-se àquelas que apresentam característica de longa dependência, ou seja, séries temporais nas quais a dependência entre observações distantes não é desprezível. Neste trabalho, analisamos o modelo ARFIN!A(p, d,q ), para dE (0,0;0,5), que apresenta a. característica de longa dependência. Como estimativas para o grau de diferenciação d consideramos os estimadores obtidos através da função periodograma, da função periodograma suavizado e da função de máxima verossimilhança sugerida por Whittle, comparando a variância e o erro quadrático médio destes estimadores através de diversas simulações.Recent work on time series analysis is concerned with the property of long mcmory, that is, time series in which the dependence between distant observations is not negligible. In this work we analyzc the ARF I .NI A(p, d, q) model, for d E (0.0; 0.5), that has the property of long memory. We consider estimators for the degree of differencing d based on the perioclogram function, on the smoothed periodogram function , anel on the maximum likelihood function suggested by Whittle. Through several simulations we compare the variance anel the mean squared error for these estimators.
9
Isbister, Tim. "Anomaly detection on social media using ARIMA models." Thesis, Uppsala universitet, Institutionen för informationsteknologi, 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-269189.
Full textAbstract:
This thesis explores whether it is possible to capture communication patterns from web-forums and detect anomalous user behaviour. Data from individuals on web-forums can be downloaded using web-crawlers, and tools as LIWC can make the data meaningful. If user data can be distinguished from white noise, statistical models such as ARIMA can be parametrized to identify the underlying structure and forecast data. It turned out that if enough data is captured, ARIMA models could suggest underlying patterns, therefore anomalous data can be identified. The anomalous data might suggest a change in the users' behaviour.
10
Cardoso, Neto Jose. "Agregação temporal de variavel fluxo em modelos Arima." [s.n.], 1990. http://repositorio.unicamp.br/jspui/handle/REPOSIP/305854.
Full textAbstract:
Orientador : Luiz Koodi HottaDissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Ciencia da Computação
Made available in DSpace on 2018-07-14T02:03:40Z (GMT). No. of bitstreams: 1 CardosoNeto_Jose_M.pdf: 1451231 bytes, checksum: 825d0beda95d7e2c6988f58b7c49500a (MD5) Previous issue date: 1990
Resumo: Não informado
Abstract: Not informed
Mestrado
Mestre em Estatística
11
Landström, Johan, and Patric Linderoth. "Precisionsbaserad analys av trafikprediktion med säsongsbaserad ARIMA-modellering." Thesis, Högskolan i Borås, Akademin för bibliotek, information, pedagogik och IT, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:hb:diva-14336.
Full textAbstract:
Intelligenta Transportsystem (ITS) utgör idag en central del i arbetet att försöka höja kvaliteten i transportnätverken, genom att exempelvis ge stöd i arbetet att leda trafik i realtid och att ge trafikanter större möjlighet att ta informerade beslut gällandes sin körning. Kortsiktig prediktion av trafikdata, däribland trafikvolym, spelar en central roll för de tjänster ITS-systemen levererar. Den starka teknologiska utvecklingen de senaste decennierna har bidragit till en ökad möjlighet till att använda datadriven modellering för att utföra kortsiktiga prediktioner av trafikdata. Säsongsbaserad ARIMA (SARIMA) är en av de vanligaste datadrivna modellerna för modellering och predicering av trafikdata, vilken använder mönster i historisk data för att predicera framtida värden. Vid modellering med SARIMA behöver en mängd beslut tas gällandes de data som används till modelleringen. Exempel på sådana beslut är hur stor mängd träningsdata som ska användas, vilka dagar som ska ingå i träningsmängden och vilket aggregationsintervall som ska användas. Därtill utförs nästintill enbart enstegsprediktioner i tidigare studier av SARIMA-modellering av trafikdata, trots att modellen stödjer predicering av flera steg in i framtiden. Besluten gällandes de parametrar som nämnts saknar ofta teoretisk motivering i tidigare studier, samtidigt som det är högst troligt att dessa beslut påverkar träffsäkerheten i prediktionerna. Därför syftar den här studien till att utföra en känslighetsanalys av dessa parametrar, för att undersöka hur olika värden påverkar precisionen vid prediktion av trafikvolym. I studien utvecklades en modell, med vilken data kunde importeras, preprocesseras och sedan modelleras med hjälp av SARIMA. Studien använde trafikvolymdata som insamlats under januari och februari 2014, med hjälp av kameror placerade på riksväg 40 i utkanten av Göteborg. Efter differentiering av data används såväl autokorrelations- och partiell autokorrelationsgrafer som informationskriterier för att definiera lämpliga SARIMA-modeller, med vilka prediktioner kunde göras. Med definierade modeller genomfördes ett experiment, där åtta unika scenarion testades för att undersöka hur prediktionsprecisionen av trafikvolym påverkades av olika mängder träningsdata, vilka dagar som ingick i träningsdata, längden på aggregationsintervallen och hur många tidssteg in i framtiden som predicerades. För utvärdering av träffsäkerheten i prediktionerna användes MAPE, RMSE och MAE. Resultaten som experimentet visar är att definierade SARIMA-modeller klarar att predicera aktuell data med god precision oavsett vilka värden som sattes för de variabler som studerades. Resultaten visade dock indikationer på att en träningsvolym omfattande fem dagar kan generera en modell som ger mer träffsäkra prediktioner än när volymer om 15 eller 30 dagar används, något som kan ha stor praktisk betydelse vid realtidsanalys. Därtill indikerar resultaten att samtliga veckodagar bör ingå i träningsdatasetet när dygnsvis säsongslängd används, att SARIMA-modelleringen hanterar aggregationsintervall om 60 minuter bättre än 30 eller 15 minuter samt att enstegsprediktioner är mer träffsäkra än när horisonter om en eller två dagar används. Studien har enbart fokuserat på inverkan av de fyra parametrarna var för sig och inte om en kombinerad effekt finns att hitta. Det är något som föreslås för framtida studier, liksom att vidare utreda huruvida en mindre träningsvolym kan fortsätta att generera mer träffsäkra prediktioner även för andra perioder under året.Intelligent Transport Systems (ITS) today are a key part of the effort to try to improve the quality of transport networks, for example by supporting the real-time traffic management and giving road users greater opportunity to take informed decisions regarding their driving. Short-term prediction of traffic data, including traffic volume, plays a central role in the services delivered by ITS systems. The strong technological development has contributed to an increased opportunity to use data-driven modeling to perform short-term predictions of traffic data. Seasonal ARIMA (SARIMA) is one of the most common models for modeling and predicting traffic data, which uses patterns in historical data to predict future values. When modeling with SARIMA, a variety of decisions are required regarding he data used. Examples of such decisions are the amount of training data to be used, the days to be included in training data and the aggregation interval to be used. In addition, one-step predictions are performed most often in previous studies of SARIMA modeling of traffic data, although the model supports multi-step prediction into the future. Often, in previous studies, decisions are made concerning mentioned variables without theoretical motivation, while it is highly probable that these decisions affect the accuracy of the predictions. Therefore, this study aims at performing a sensitivity analysis of these parameters to investigate how different values affect the accuracy of traffic volume prediction. The study developed a model with which data could be imported, preprocessed and then modeled using a SARIMA model. Traffic volume data was used, which was collected during January and February 2014, using cameras located on highway 40 on the outskirts of Gothenburg. After differentiation of data, autocorrelation and partial autocorrelation graphs as well as information criteria are used to define appropriate SARIMA models, with which predictions could be made. With defined models, an experiment was conducted in which eight unique scenarios were tested to investigate how the prediction accuracy of traffic volume was influenced by different amount of exercise data, what days was included in training data, length of aggregation intervals, and how many steps into the future were predicted. To evaluate the accuracy of the predictions, MAPE, RMSE and MAE were used. The results of the experiment show that developed SARIMA models are able to predict current data with good precision no matter what values were set for the variables studied. However, the results showed indications that a training volume of five days can generate a model that provides more accurate predictions than when using 15 or 30-day volumes, which can be of great practical importance in real-time analysis. In addition, the results indicate that all weekdays should be included in the training data set when daily seasonality is used, SARIMA modeling handles aggregation intervals of 60 minutes better than 30 or 15 minutes, and that one-step predictions are more accurate than when one or two days horizons are used. The study has focused only on the impact of the four parameters separately and not if a combined effect could be found. Further research is proposed for investigating if combined effects could be found, as well as further investigating whether a lesser training volume can continue to generate more accurate predictions even for other periods of the year.
12
Nayeri, Negin. "Option strategies using hybrid Support Vector Regression - ARIMA." Thesis, KTH, Matematisk statistik, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-275719.
Full textAbstract:
In this thesis, the use of machine learning in option strategies is evaluated with focus on the S&P 500 Index. The first part of the thesis focuses on testing the performance power of the Support Vector Regression (SVR) method for the historical realized volatility with a window of 20 days. The prediction window will also be 1-month forward (approximately 20 trading days). The second part of the thesis focuses on creating an ARIMA model that forecasts the error that is based on the difference between the predicted respective true values. This is done in order to create the hybrid SVR-ARIMA model. The new model now consists of a realized volatility value derived from the SVR model as well as the error obtained from the ARIMA model. Lastly, the two methods, that is single SVR and hybrid SVR-ARIMA are compared and the model that exhibits the best result is used within two option strategies. The results showcase the promising forecasting power of the SVR method which for this dataset had an accuracy leveland 67 % for the realized volatility. The ARIMA model also exhibits successful forecasting ability for the next lag. However, for this dataset, the Hybrid SVR-ARIMA model outperforms the single SVR model. It is debatable whether the success of these methods may be due to the fact the dataset only covers the years between 2010-2018 and the highly volatile environments of the financial crisis 2008 is omitted. Nonetheless, the use of the hybrid SVR-ARIMA model used within the two option strategies gives an average payoff 0.37 % and 1.68 %. It should however be noted that the affiliated costs of trading options is not included in the payoff and neither is the cost of premium due in buying options as the costs vary depending on the origin of the purchase. This thesis has been performed in collaboration with Crescit Asset Management in Stockholm, Sweden.I denna uppsats utvärderas användningen av maskininlärning i optionsstrategier med fokus på S&P 500 Index. Den första delen av uppsatsen fokuserar på att testa prognos kraften av Support Vector Regression (SVR) metoden för den realiserade volatiliteten med ett fönster på 20 dagar. Prognos kommer att ske för 1 månad framåt (20 trading dagar). Den andra delen av uppsatsen fokuserar på att skapa en ARIMA-modell som prognostiserar nästa värdet i tidsserien som baseras på skillnaden mellan de erhållna prognoserna samt sanna värdena. Detta görs för att skapa en hybrid SVR-ARIMA-modell. Den nya modellen består nu av ett realiserat volatilitetsvärde härrörande från SVR samt den error som erhållits från ARIMA- modellen. Avslutningsvis kommer de två metoderna, det vill säga SVR och hybrid SVR-ARIMA, jämföras och den modell med bäst resultat användas inom två options strategier. Resultaten visar den lovande prognotiseringsförmågan för SVR-metoden som för denna dataset hade en noggrannhetsnivå på 67 % för realiserad volatiliteten. ARIMA- modellen visar också en framgångsrik prognosförmåga för nästa punkt i tidsserien. Dock överträffar Hybrid SVR-ARIMA-modellen SVR-modellen för detta dataset. Det kan diskuteras ifall framgången med dessa metoder kan bero på att denna dataset täcker åren mellan 2010-2018 och det mycket volatila tiden under finanskrisen 2008 är uteslutet. Detta kan ifrågasätta modellernas prognotiseringsförmåga på högre volatilitetsmarknader. Dock ger användningen av hybrid-SVR-ARIMA-modellen som används inom de två option strategierna en genomsnittlig avkastning på 0,37 % och 1,68 %. Det bör dock noteras att de tillkommande kostnaderna för att handla optioner samt premiekostnaden som skall betalas i samband med köp av optioner inte ingår i avkastningen då dessa kostnader varierar beroende på plats av köp. Denna uppsats har gjorts i samarbete med Crescit Asset Management i Stockholm.
13
Wenzel, Anne. "Komponentenzerlegung des Regelleistungsbedarfs mit Methoden der Zeitreihenanalyse." Master's thesis, Universitätsbibliothek Chemnitz, 2011. http://nbn-resolving.de/urn:nbn:de:bsz:ch1-qucosa-66420.
Full textAbstract:
Im Rahmen der Arbeit wurden die minutengenauen Daten des Regelleistungsbedarfs (Summe aus Sekundärregelleistung und Minutenreserve) der Monate April bis Dezember des Jahres 2009 einer Regelzone einer Zeitreihenanalyse unterzogen und in Komponenten gemäß dem klassischen Komponentenmodell zerlegt. Diese sind die Trendkomponente, ermittelt durch einen gleitenden Durchschnitt mit der Länge einer Stunde, weiterhin zwei periodische Komponenten mit der Periodenlänge einer Stunde sowie der Periodenlänge eines Tages und die Restkomponente, welche mit einem ARIMA(2,1,5)-Prozess modelliert wurde. In der Zukunft sollte das erstellte Modell des Regelleistungsbedarfs durch Hinzunahme einer jahreszeitlichen Komponente noch verbessert werden. Dies war im Rahmen der Arbeit nicht möglich, da keine Daten über einen Zeitraum von mehreren Jahren vorhanden waren. Zusätzlich kann geprüft werden, inwiefern mit dem Komponentenmodell Prognosen durchführbar sind. Dafür sollte die Trendkomponente anders gewählt werden, da sich der hier gewählte Weg zu sehr an den Daten orientiert. Der zweite Teil der Aufgabenstellung dieser Arbeit bestand im Identifizieren inhaltlicher Komponenten, also möglicher Zusammenhänge zwischen dem Regelleistungsbedarf und verschiedenen denkbaren Ursachen. Als potentielle Ursachen wurden der Lastverlauf sowie die Windenergieeinspeisung untersucht. Zwischen der Zeitreihe des Lastverlaufs und der des Regelleistungsbedarfs bestand eine leichte positive Korrelation, zwischen der Zeitreihe der Windenergieeinspeisung und der des Regelleistungsbedarfs eine geringe negative Korrelation.
14
Andréasson, David, and Blomquist Jesper Mortensen. "Forecasting the OMXS30 - a comparison between ARIMA and LSTM." Thesis, Uppsala universitet, Statistiska institutionen, 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-413793.
Full textAbstract:
Machine learning is a rapidly growing field with more and more applications being proposed every year, including but not limited to the financial sector. In this thesis, historical adjusted closing prices from the OMXS30 index are used to forecast the corresponding future values using two different approaches; one using an ARIMA model and the other using an LSTM neural network. The forecasts are made on three different time intervals: 90, 30 and 7 days ahead. The results showed that the LSTM model performs slightly better when forecasting 90 and 30 days ahead, whereas the ARIMA model has comparable accuracy on the seven day forecast.
15
Koliadenko, Pavlo <1998>. "Time series forecasting using hybrid ARIMA and ANN models." Master's Degree Thesis, Università Ca' Foscari Venezia, 2021. http://hdl.handle.net/10579/19992.
Full text
16
Uppling, Hugo, and Adam Eriksson. "Single and multiple step forecasting of solar power production: applying and evaluating potential models." Thesis, Uppsala universitet, Institutionen för teknikvetenskaper, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-384340.
Full textAbstract:
The aim of this thesis is to apply and evaluate potential forecasting models for solar power production, based on data from a photovoltaic facility in Sala, Sweden. The thesis evaluates single step forecasting models as well as multiple step forecasting models, where the three compared models for single step forecasting are persistence, autoregressive integrated moving average (ARIMA) and ARIMAX. ARIMAX is an ARIMA model that also takes exogenous predictors in consideration. In this thesis the evaluated exogenous predictor is wind speed. The two compared multiple step models are multiple step persistence and the Gaussian process (GP). Root mean squared error (RMSE) is used as the measurement of evaluation and thus determining the accuracy of the models. Results show that the ARIMAX models performed most accurate in every simulation of the single step models implementation, which implies that adding the exogenous predictor wind speed increases the accuracy. However, the accuracy only increased by 0.04% at most, which is determined as a minimal amount. Moreover, the results show that the GP model was 3% more accurate than the multiple step persistence; however, the GP model could be further developed by adding more training data or exogenous variables to the model.
17
Zatloukal, Radomír. "Analýza a předpověď časových řad pomocí statistických metod se zaměřením na metodu Box-Jenkins." Master's thesis, Vysoké učení technické v Brně. Fakulta strojního inženýrství, 2008. http://www.nusl.cz/ntk/nusl-228167.
Full textAbstract:
Two real time series, one discussing the area of energy, other discussing the area of economy. By the energetic area we will be dealing with the electric power consumption in the USA, by the economic area we will be dealing with the progress of index PX50. We will try to approve the validity of hypothesis that with some test functions we will be able to set down the accidental unit distribution in these two time series.
18
Bahri, El Mostafa. "L'identification automatique des processus ARIMA : une approche par système expert." Aix-Marseille 3, 1991. http://www.theses.fr/1991AIX32043.
Full textAbstract:
L'approche arima de prevision des series chronologiques, mise au point en 1970 par box et jenkins, s'est averee tres pertinente dans sa philosophie et tres satisfaisante au niveau des resultats dans la pratique, et ce en comparaison avec d'autres methodes du meme propos. Cependant, l'indentification des processus arima reste une tache hors de portee des non-specialistes en la matiere, a cause notamment de sa nature heuristique. Ceci explique pourquoi cette approche n'a pas suffisamment penetre les milieux des utilisateurs des methodes de previsions. Aussi, pour cette meme raison, l'automatisation de cette etape centrale de la methodologie arima est inefficace via la voie purement procedurale de l'informatique. Notre propos est que l'automatisation par la technique des systemes experts est mieux adaptee. La premiere partie de ce travail se propose d'argumenter cette these, a la lumiere des particularites du probleme de l'indentification des processus arima, et a travers la litterature la plus recente consacree a cette question. Nous avons aussi entrepris des comparaisons empiriques dans cette direction. Dans la seconde partie, nous avons montre l'interet de l'approche par systeme expert. Puis concu un prototype de systeme expert d'identification des processus arima. Nous avons realise ce prototype, en langage a base de regles vax-ops5, au sein du groupe d'intelligence artificielle de digital equipement (sophia-antipolis). Enfin, a travers cette application, intervient une evaluation de la methodologie des systemes experts dans le domaine du traitement automatique des series chronologiques, et des orienArima approach is an important contribution in fore casting economic time series but indentifying such processes is a crucial task, both manualy ans automatically we suggest that the expert system approach is an adequate solution for this problem. We have written a prototype in poss for this purpose and we propose neural network as complementary technique for automatic identification of series procecesses
19
Heed, Ingrid, and Karl Lindberg. "Forecasting COVID-19 hospitalizations using dynamic regression with ARIMA errors." Thesis, Uppsala universitet, Statistiska institutionen, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-446310.
Full textAbstract:
For more than a year, COVID-19 has changed societies all over the world and put massive strains on its healthcare systems. In an attempt to aid in prioritizing medical resources, this thesis uses dynamic regression with ARIMA errors to forecast the number of hospitalizations related to COVID-19 two weeks ahead in Uppsala County. For this purpose, 100 models are created and their ability to forecast hospitalizations two weeks ahead for weeks 15-17 of 2021 for the different municipalities in Uppsala County is evaluated using root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). The best performing models are then utilized to forecast hospitalizations for weeks 19-22. The results show that the models perform well during periods of increasing numbers of hospitalizations during early 2021, while they perform less well during the last weeks of May 2021 where hospitalizations numbers have been falling dramatically. This recent decrease in forecasting performance is believed to be caused by an increase in vaccination coverage, which is not accounted for in the models.
20
Urettini, Edoardo <1997>. "Combination of forecasts from ARIMA, Neural Networks and Hybrid models." Master's Degree Thesis, Università Ca' Foscari Venezia, 2021. http://hdl.handle.net/10579/19877.
Full textAbstract:
Lo scopo della tesi è mostrare se una combinazione di previsioni di diversi tipi di modelli possa migliorare le capacità predittive rispetto ai modelli presi separatamente. Sono state utilizzate tre diverse classi di modelli: modelli ARIMA-GARCH, reti neurali e una ibridazione tra queste due classi. La combinazione delle previsioni di queste diverse classi cerca di estrarne le capacità uniche nello spiegare una serie storica, andando oltre la generalizzazione fornita da un unico modello ibrido. Viene presentata una applicazione sulla previsione dell'indice VIX.
21
Silva, Alyne Neves. "Detecção de outliers em séries espaço-temporais: análise de precipitação em Minas Gerais." Universidade Federal de Viçosa, 2012. http://locus.ufv.br/handle/123456789/4061.
Full textAbstract:
Made available in DSpace on 2015-03-26T13:32:17Z (GMT). No. of bitstreams: 1
texto completo.pdf: 3004404 bytes, checksum: 18834db766750ae443a52c29a9b0decd (MD5)
Previous issue date: 2012-07-24Fundação de Amparo a Pesquisa do Estado de Minas Gerais
Time series are sometimes influenced by disruptions of events, such as strikes, the outbreak of war, among others. These interrupts originate atypical observations or outliers that directly influence the homogeneity of the series, leading to erroneous inferences and interpretations of the variable under study, being very common in climatological data. So, in the interest of detecting outliers in time series of precipitation, this study aimed to establish a method of detecting outliers. For this, there was the junction of ARIMA models and methodologies of the classical geostatistics, the self-validation. The proposed criterion compares waste of time series analysis with confidence intervals of the residue of self-validation. We analyzed time series of average monthly rainfall for rainy days of 43 rainfall stations in the state of Minas Gerais, between the years 2000 to 2005. The analysis procedures ranging from the description of the periodicity through the periodogram to obtain validation, from the estimation of the semivariogram models by ordinary least squares methods and maximum likelihood. The results for the period under study, 165 were detected outliers, spread between the 43 rainfall stations. The station Campo Grande Ranch, located in the municipality of Passa Tempo, was the season in which they recorded the highest number of outliers, 45 in total. As the results, we considered the proposed method very efficient in detecting outliers, and therefore the analysis of the homogeneity of observations.
Séries temporais são algumas vezes influenciadas por interrupções de eventos, tais como greves, eclosão de guerras, entre outras. Estas interrupções originam observações atípicas ou outliers que influenciam diretamente na homogeneidade da série, ocasionando interpretações e inferências errôneas da variável sob estudo, sendo muito comum em dados climatológicos. Assim, com o interesse de detectar outliers em séries temporais de precipitação, o presente trabalho teve por objetivo estabelecer um método de detecção outliers. Para tal, realizou-se a junção da modelagem ARIMA e de uma das metodologias clássicas de geoestatística, a autovalidação. O critério proposto compara os resíduos da análise de séries temporais com intervalos de confiança dos resíduos da autovalidação. Foram analisadas séries temporais da precipitação média mensal por dias chuvosos de 43 estações pluviométricas localizadas no estado de Minas Gerais, entre os anos de 2000 a 2005. Os procedimentos de análise vão da descrição da periodicidade por meio do periodograma até a obtenção da autovalidação, à partir da estimação dos modelos de semivariograma pelos métodos de mínimos quadrados ordinários e máxima verossimilhança. Pelos resultados, para o período sob estudo, foram detectado 165 outliers, espalhados entre as 43 estações pluviométricas. A estação Fazenda Campo Grande, localizada no município de Passa Tempo, foi a estação em que se registrou o maior número de outliers, 45 no total. Conforme os resultados obtidos considerou-se o método proposto muito eficiente na detecção de outliers e, consequentemente, na análise da homogeneidade das observações.
22
Odencrants, Martin, and Fredrik Rahm. "Säsongsrensning : En komparativ studie av TRAMO/SEATS och X-12 ARIMA." Thesis, Örebro University, Department of Business, Economics, Statistics and Informatics, 2007. http://urn.kb.se/resolve?urn=urn:nbn:se:oru:diva-1760.
Full textAbstract:
Ett syfte med tidserieteori är att dekomponera en observerad tidsserie i en summa icke observerbara komponenter. Dessa komponenter är Trend, Cykel, Säsong, Kalendereffekter, Extremvärden samt Irreguljära effekter.
Det finns två olika teorier för dekomponering av tidsserier, modellbaserad dekomponering och icke modellbaserad dekomponering. De två olika teorierna skiljer sig åt i grunden. Den här uppsatsen syftar till att utvärdera de två säsongsrensningsmetoderna TRAMO/SEATS och X-12 ARIMA samt att säsongsrensa tidsserien över den totala lönesumman, vilken är en del av statistikprodukten Lönesummor arbetsgivaravgifter och preliminär A-skatt (LAPS) producerad av SCB.
23
Gustavsson, André. "Elpriserna på den nordiska elbörsen : Prognosmodellering med hjälp av ARIMA-modeller." Thesis, Umeå University, Department of Statistics, 2009. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-34820.
Full text
24
Holens, Gordon Anthony. "Forecasting and selling futures using ARIMA models and a neural network." Thesis, National Library of Canada = Bibliothèque nationale du Canada, 1997. http://www.collectionscanada.ca/obj/s4/f2/dsk3/ftp05/mq23343.pdf.
Full text
25
Putzulu, Matteo. "Modelli ARIMA implementati in ambiente Python applicati a serie temporali GNSS." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2022. http://amslaurea.unibo.it/25884/.
Full textAbstract:
Nella presente tesi, si è discusso sul corretto trattamento dei dati di posizione, provenienti da una stazione permanente GPS in PPP, per studiarne l’andamento e successivamente elaborarne le previsioni per il futuro. E' stato utlizzato un approccio con la classe del Modelli ARIMA implementati su linguaggio Python.
26
Akonom, Jacques. "Processus transformés d'un ARIMA ou d'un processus de Wiener : Problèmes d'estimation." Lille 1, 1988. http://www.theses.fr/1988LIL10112.
Full textAbstract:
Utilisant un resultat de komlos major tusnady sur l'approximation forte des sommes partielles d'une suite de variables aleatoires independantes identiquement distribuefes, on etablit la convergence presque sure du temps d'occupation d'un intervale par une marche aleatoire vers celui du mouvement brownien. Dans le second chapitre, on etudie les sommes partielles d'un processus lineaire stationnaire et on donne des conditions d'approximation forte de ce processus par un processus de wiener. On en deduit la convergence en loi du temps d'occufpation d'un intervale par un processus arima d'ordre 1. Le chapitre suivant est consacre a un probleme d'estimation. On etudie l'image par une application continue d'un processus arima d'ordre 1. On propose lorsqu'un tel processus est observe un estimateur de la transformation reciproque, ainsi qu'un estimateur de la fonction derivee. Enfin on etudie les processus arima fractionnaires, dans le cas non stationnaire. On discute le choix des conditions initiales et on etablit que le processus obtenu apres normalisation converge en loi vers le processus du mouvement brownien fractionnaire de b. Mandelbrot. En annexe, des resultats recents sur la melangeance des processus lineaires, et plus particulierement les arma, sont donnes
27
Akonom, Jacques. "Processus transformés d'un ARIMA ou d'un processus de Wiener problèmes d'estimation /." Grenoble 2 : ANRT, 1988. http://catalogue.bnf.fr/ark:/12148/cb376111363.
Full text
28
Ribeiro, Liliana Patrícia Teixeira. "Aplicação de modelos econométricos na previsão de preço de azeites." Master's thesis, Instituto Superior de Economia e Gestão, 2020. http://hdl.handle.net/10400.5/20862.
Full textAbstract:
Mestrado em Econometria Aplicada e PrevisãoO presente relatório tem por base as atividades desenvolvidas no estágio na empresa Gallo Worldwide, nomeadamente a análise das bases de dados da empresa de modo a efetuar a previsão do preço do azeite extra-virgem, azeite virgem e lampante. Uma vez que a modelação dos preços dos azeites é realizada através da modelação de séries temporais, existem diversos modelos que podem ser aplicados. Segundo a literatura científica analisada, a estimação das séries temporais utilizadas pode ser realizada através do modelo ARIMA, ARIMAX, GARCH e SUR. Neste sentido, será apresenta de uma forma detalhada a análise dos modelos econométricos em estudo para a obtenção das previsões pretendidas. Os modelos utilizados foram aplicados a conjuntos de dados com diferentes periodicidades: semanal e mensal. Sendo os modelos aplicados a conjuntos de dados com diferentes periodicidades também foram efetuadas previsões através de todos os modelos aplicados aos dois conjuntos de dados, existindo conclusões para ambos os casos.
The current report was built around the tasks performed during the internship on the company Gallo Worldwide, where the main responsibilities consisted in the analysis of the database to be able to forecast extra-virgin olive oil, virgin olive oil and lampante prices. Considering the olive oil pricing modelling is achieved through the modelling of time series, several models can be applied. According to the scientific literature reviewed, the estimation of time series may be accomplished using the ARIMA, ARIMAX, GARCH and SUR models. In this sense, it will be presented, in a detailed manner, the analysis of the econometrical models being studied as a resource to obtain the intended predictions. The models utilized were applied to a group of data with different periodicities: data with weekly periodicity and data with monthly periodicity. Considering the models are employed over a set of data with different periodicities, similarly the predictions were made through all the models used in both sets of data, resulting in the existence ofconclusions for both cases.
info:eu-repo/semantics/publishedVersion
29
Cruz, Cristovam Colombo dos Santos. "AnÃlise de sÃries temporais para previsÃo mensal do icms: o caso do PiauÃ." Universidade Federal do CearÃ, 2007. http://www.teses.ufc.br/tde_busca/arquivo.php?codArquivo=1648.
Full textAbstract:
nÃo hÃEsta DissertaÃÃo trata de pesquisa sobre a anÃlise de sÃries temporais para previsÃo mensal do Imposto Sobre CirculaÃÃo e Mercadorias e PrestaÃÃo de ServiÃos â ICMS no estado do PiauÃ. Objetiva-se com essa pesquisa oferecer aos gestores do estado um modelo de previsÃo consistente e com bom poder preditivo, de forma a contribuir com a gestÃo financeira estadual. No trabalho, utilizaram-se os modelos ARIMA e FunÃÃo de TransferÃncia para realizar previsÃes, bem como o Modelo CombinaÃÃo de PrevisÃes. A dissertaÃÃo apresenta um diagnÃstico do ICMS no estado do Piauà e uma revisÃo da literatura onde sÃo abordados os principais aspectos teÃricos dos modelos utilizados no trabalho, bem como a anÃlise dos resultados empÃricos. Ao final, pode-se observar que os resultados obtidos na presente dissertaÃÃo, estÃo em sintonia com outros resultados obtidos em trabalhos semelhantes realizados sobre o tema, o que vem a confirmar a importÃncia dos modelos que utilizam a anÃlise de sÃries temporais como instrumento de prediÃÃo.
This dissertation deals with a research on the temporal series analysis for the monthly forecast of the turnover and services tax â ICMS in Brazil â in the state of PiauÃ. The aim of this research is to offer the statewide policymakers a consistent forecast and powerfully predictive model, so as to contribute to the state finance management. In this work, the ARIMA and Assignment Function models were used to carry out forecasts, as well as Forecast Combination. The dissertation presents a diagnosis of the ICMS in the state of PiauÃ, a review on the literature where the main theoretical aspects of the models carried out in the work are addressed, in addition to the empirical findings analysis. As a conclusion, it can be observed that the findings carried out in this dissertation are in harmony with other results of similar works carried out on the theme, which corroborates the importance of the models using the temporal series analysis as a forecasting instrument.
30
Becker, Claudia. "Die Re-Analyse von Monitor-Schwellenwerten und die Entwicklung ARIMA-basierter Monitore für die exponentielle Glättung /." Aachen : Shaker, 2006. http://www.gbv.de/dms/zbw/51982640X.pdf.
Full text
31
Mohamed, Fadil B. "Space-time ARIMA and transfer function-noise modeling of rainfall-runoff process." Thesis, University of Ottawa (Canada), 1985. http://hdl.handle.net/10393/4723.
Full text
32
Almeida, Silvana Gonçalves de. "ANÁLISE DO CUSTO DE MEDICAMENTOS QUIMIOTERÁPICOS, POR MEIO DE MODELOS ARIMA - ARCH." Universidade Federal de Santa Maria, 2011. http://repositorio.ufsm.br/handle/1/8196.
Full textAbstract:
Coordenação de Aperfeiçoamento de Pessoal de Nível SuperiorToday's medical treatments are becoming more expensive, in view of this plan and control costs are mechanisms that can ensure the survival of hospitals. The present study analyzed the cost of medications relevant financial, between January 2003 and November 2010, at University Hospital of Santa Maria. Since not all items have the same degree of importance, the drugs were classified by the ABC method which provided work with capecitabine and imatinib, the total cost of these drugs in 2010, representing about 18% compared to total expenditure on drugs and materials. The models found for the series of the cost of capecitabine and imatinib were ARIMA (0,1,1)-ARCH (1) and ARIMA (1,1,0)-ARCH (1), respectively. These models were used to analyze the behavior of the series under study and make predictions in order to assist hospital managers in decision making in hospital inventory management.
Atualmente os tratamentos médicos estão cada vez mais caros, em vista disso planejar e controlar custos são mecanismos que podem garantir a sobrevivência das instituições hospitalares. O presente estudo analisou o custo com medicamentos de relevância financeira, entre janeiro de 2003 e novembro de 2010, no Hospital Universitário de Santa Maria. Como nem todos os itens têm o mesmo grau de importância, os medicamentos foram classificados pelo método ABC o que proporcionou trabalhar com a Imatinibe e Capecitabina, cujo custo total em 2010 destes medicamentos, representou cerca de 18% em relação ao gasto total com medicamentos e materiais. Os modelos encontrados para as séries do custo de Imatinibe e Capecitabina foram, ARIMA(0,1,1)-ARCH(1) e ARIMA(1,1,0)-ARCH(1), respectivamente. Tais modelos foram utilizados para analisar o comportamento das séries em estudo e realizar previsões com o objetivo de auxiliar os gestores hospitalares nas tomadas de decisões no gerenciamento de estoque hospitalar.
33
Elmasdotter, Ajla, and Carl Nyströmer. "A comparative study between LSTM and ARIMA for sales forecasting in retail." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-229747.
Full textAbstract:
Food waste is a major environmental issue. Expired products are thrown away, implying that too much food is ordered compared to what is sold and that a more accurate prediction model is required within grocery stores. In this study the two prediction models Long Short-Term Memory (LSTM) and Autoregressive Integrated Moving Average (ARIMA) were compared on their prediction accuracy in two scenarios, given sales data for different products, to observe if LSTM is a model that can compete against the ARIMA model in the field of sales forecasting in retail. In the first scenario the models predict sales for one day ahead using given data, while they in the second scenario predict each day for a week ahead. Using the evaluation measures RMSE and MAE together with a t-test the results show that the difference between the LSTM and ARIMA model is not of statistical significance in the scenario of predicting one day ahead. However when predicting seven days ahead, the results show that there is a statistical significance in the difference indicating that the LSTM model has higher accuracy. This study therefore concludes that the LSTM model is promising in the field of sales forecasting in retail and able to compete against the ARIMA model.Matsvinn är ett stort problem för miljön. Utgångna produkter slängs, vilket implicerar att för mycket mat beställs jämfört med hur mycket butikerna säljer. En mer precis modell för att förutsäga försäljningssiffrorna kan minska matsvinnet. Denna studie jämför modellerna Long Short-Term Memory (LSTM) och Autoregressive Integrated Moving Average (ARIMA) i deras precision i två scenarion. Givet försäljningssiffror för olika matvaruprodukter, undersöks ifall LSTM är en modell som kan konkurrera mot ARIMA-modellen när modellerna ska förutsäga försäljningssiffror för matvaruprodukter. Det första scenariot var att förutse försäljningen en dag i framtiden baserat på given data, medan det andra scenariot var att förutse försäljningen varje dag under en vecka i framtiden baserat på given data. Genom att använda måtten RMSE och MAE tillsammans med ett T-Test visade resultaten av studien att skillnaden mellan LSTM- och ARIMA-modellen inte var av statistik signifikans i fallet då modellerna skulle förutsäga försäljningen en dag i framtiden. Däremot visar resultaten på att skillnaden mellan modellerna är av signifikans när modellerna skulle förutsäga försäljningen under en vecka, vilken implicerar att LSTM-modellen har en högre precision i detta scenario. Denna studie drar därmed slutsatsen att LSTM-modellen är lovande och kan konkurrera mot ARIMA-modellen när det kommer till försäljningssiffror av matvaruprodukter.
34
Borneklint, Niklas. "Forecasting prices of Bitcoin and Google stock with ARIMA vs Facebook Prophet." Thesis, Högskolan Väst, Avd för juridik, ekonomi, statistik och politik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:hv:diva-17345.
Full textAbstract:
In this thesis we have presented econometrics and forecasts of Bitcoin and Google (GOOG) prices. We have implemented two models, one traditional, “ARIMA” and a relatively new one, “Prophet model” by using Facebook Prophet (ML). Machine learning is still new in the economic field, it has been rewarding to learn its capability. We have evaluated the model’s performance by using root mean square error (RMSE) and compared the result which model performed better. We wanted to compare to different assets, volatile Bitcoin to considerable stable Google (GOOG), thus investigate our models performance and if they differ or not. Regarding our result, we found that the ARIMA models have the best forecasting ability. We also investigate the impact of rational expectation and its impact on an asset price. We found that announcements on Bitcoin cause a significantly change in price and had an impact on the model’s performance.I denna avhandling har vi presenterat ekonometriska modeller och prognoserade prisnivåer av Bitcoins och Googles (GOOG). Vi har implementerat två modeller, en traditionell, "ARIMA" samt en relativt ny modell, "Profetmodellen" med Facebook Prophet (ML). Maskininlärning är fortfarande nytt inom det ekonomiska området och det har varit givande att förstå dess förmåga. Vi vill jämföra två typer av tillgångar, Bitcoin som är volatile mot Google som är förhållandevis stabil för att se om våra modeller skiljer sig åt. Vi har utvärderat modellens prestanda med hjälp av root mean square error (RMSE) och jämförde resultatet vilken modell som var mest exakt. Vi fann att ARIMA-modellen gav oss bäst resultat. Vi undersöker också effekterna av rationella förväntningar och dess inverkan på pris av tillgång. Vi fann att nyheter om Bitcoin influerar dess pris och hade en inverkan på modellernas prestanda.
35
Wågberg, Max. "Att förutspå Sveriges bistånd : En jämförelse mellan Support Vector Regression och ARIMA." Thesis, Mittuniversitetet, Institutionen för informationssystem och –teknologi, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:miun:diva-36479.
Full textAbstract:
In recent years, the use of machine learning has increased significantly. Its uses range from making the everyday life easier with voice-guided smart devices to image recognition, or predicting the stock market. Predicting economic values has long been possible by using methods other than machine learning, such as statistical algorithms. These algorithms and machine learning models use time series, which is a set of data points observed constantly over a given time interval, in order to predict data points beyond the original time series. But which of these methods gives the best results? The overall purpose of this project is to predict Sweden’s aid curve using the machine learning model Support Vector Regression and the classic statistical algorithm autoregressive integrated moving average which is abbreviated ARIMA. The time series used in the prediction are annual summaries of Sweden’s total aid to the world from openaid.se since 1998 and up to 2019. SVR and ARIMA are implemented in python with the help of the Scikit- and Statsmodels libraries. The results from SVR and ARIMA are measured in comparison with the original value and their predicted values, while the accuracy is measured in Root Square Mean Error and presented in the results chapter. The result shows that SVR with the RBF-kernel is the algorithm that provides the best results for the data series. All predictions beyond the times series are then visually presented on a openaid prototype page using D3.jsUnder det senaste åren har användningen av maskininlärning ökat markant. Dess användningsområden varierar mellan allt från att göra vardagen lättare med röststyrda smarta enheter till bildigenkänning eller att förutspå börsvärden. Att förutspå ekonomiska värden har länge varit möjligt med hjälp av andra metoder än maskininlärning, såsom exempel statistiska algoritmer. Dessa algoritmer och maskininlärningsmodeller använder tidsserier, vilket är en samling datapunkter observerade konstant över en given tidsintervall, för att kunna förutspå datapunkter bortom den originella tidsserien. Men vilken av dessa metoder ger bäst resultat? Projektets övergripande syfte är att förutse sveriges biståndskurva med hjälp av maskininlärningsmodellen Support Vector Regression och den klassiska statistiska algoritmen autoregressive integrated moving average som förkortas ARIMA. Tidsserien som används vid förutsägelsen är årliga summeringar av biståndet från openaid.se sedan år 1998 och fram till 2019. SVR och ARIMA implementeras i python med hjälp av Scikit-learn och Statsmodelsbiblioteken. Resultatet från SVR och ARIMA mäts i jämförelse mellan det originala värdet och deras förutspådda värden medan noggrannheten mäts i root square mean error och presenteras under resultatkapitlet. Resultatet visar att SVR med RBF kärnan är den algoritm som ger det bästa testresultatet för dataserien. Alla förutsägelser bortom tidsserien presenteras därefter visuellt på en openaid prototypsida med hjälp av D3.js.
36
Wilczek, Andrej, and Oskar Erlandsson. "Evaluering av LASSO och ARIMA algoritmerna för prognostisering i den finansiella marknaden." Thesis, KTH, Skolan för teknikvetenskap (SCI), 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-255711.
Full textAbstract:
Att förutspå händelser i aktiemarknaden anses vara en särskilt utmanande uppgift på grund av dess komplexitet och volatilitet. I detta projekt utvärderar vi befintliga maskininlärningsalgoritmer som metoder för modellering och prognostisering i finansmarknaden. I vårt försök att förutspå stängningsvärdet på Nestlés aktiekurs, implementerades linjära LASSO- och ARIMA-modeller baserat på antagandet att datat har ett linjärt beroende. Metoderna utvärderades sedan genom att beräkna tre stycken feltermer baserat på metodernas prestanda gällande kortsiktiga och långsiktiga förutsägelser. Våra resultat tyder på att LASSO-algoritmen fungerar bättre med avseende på kortsiktiga förutsägelser medan ARIMA ger mer exakta långsiktiga förutsägelser. När det gäller förutsägelse av framtida trender visar båda metoderna god övergripande prestanda. Slutligen föreslår vi intressanta områden att överväga för att kunna göra mer precisa förutsägelser när data av hög volatilitet används.Stock market forecasting is considered to be a particularly challenging task due to the complexity and volatility of the stock market. In this project we evaluate the performance of existing machine learning techniques as methods for modeling and predicting patterns in the financial market. In our attempt to predict the Nestl\'e stock closing price point, linear LASSO and ARIMA models were implemented based on the assumption that the volatile data has some type of linear dependency. The methods was evaluated by calculating the Mean Absolute Deviation, Mean Squared Error and Mean Absolute Percentage Error values based on their performance in making short and long-term predictions. Our results suggest that the LASSO algorithm performs better in regards to short-term predictions whereas the ARIMA provides more accurate long-term predictions. In terms of prediction of future trends, both methods show good overall performance. Finally, we propose interesting areas to consider in order to make more precise predictions on volatile data.
37
Teodoro, Valiana Alves. "Modelos de séries temporais para temperatura em painéis de cimento-madeira." Universidade de São Paulo, 2015. http://www.teses.usp.br/teses/disponiveis/11/11134/tde-07042015-102815/.
Full textAbstract:
Por meio do monitoramento da evolução da temperatura da mistura cimento-madeira, pode-se utilizar esta informação como uma série temporal. O objetivo deste estudo foi utilizar modelos de séries temporais para descrever as séries de temperatura do experimento constituído por diferentes espécies associadas a resíduos de Candeia na produção de painéis particulado e compara-las duas a duas para averiguar se foram geradas pelo mesmo processo estocástico. Inicialmente foi realizado um estudo para avaliar a estacionariedade das séries utilizando o correlograma e o teste da raiz unitária de Dickey-Fuller, na qual todas as séries apresentaram não estacionariedade, para o tratamento de 25% Candeia e Eucalipto com tratamento prévio de água foi dita uma série I(2) e pelos critérios AIC, BIC e MAPE o melhor modelo foi ARIMA(2, 2, 2), para o tratamento de 50% Candeia e Eucalipto também com tratamento prévio de água foi dita uma série I(1) e pelos critérios o melhor modelo foi ARIMA(4, 2, 2), para o tratamento de 75% Candeia e Eucalipto com tratamento prévio de água foi dita uma série I(1) com o modelo ARIMA(5, 1, 0), e para o tratamento de 25% Candeia e Eucalipto sem tratamento prévio de água foi dita uma série I(1) com o modelo ARIMA(2, 1, 2). Em relação à comparação das séries temporais contempladas neste trabalho é possível concluir que as mesmas são diferentes entre si, ou seja, não foram geradas pelo mesmo processo estocástico.By monitoring the temperature evolution of the cement-wood mixture, one can utilize this information as a time series. The objective of this study was to utilize time series models to describe the temperature series from an experiment, consisting of different species associated to Candeia residuals in the production of particleboard panels, and do a pairwise comparison to verify if they were generated from the same stochastic process. Initially it was realized the Dickey-Fuller unit root test to verify series stationarity, which indicated that all series were not stationary. For the 25% Candeia and Eucalyptus treatment with previous water treatment the series was best modelled by an ARIMA(2, 2, 2) as evidenced by the AIC, BIC and MAPE criteria. For the 50% Candeia and Eucalyptus treatment also with previous water treatment the series was best modelled by an ARIMA(4, 2, 2) as indicated by the same criteria. Finally for the 75% Candeia and Eucalyptus treatment with previous water treatment and the 25% Candeia and Eucalyptus treatment without previous water treatment the best models were the ARIMA(5, 1, 0) and the ARIMA(2, 1, 2) respectively. In relation to the comparison of the time series contemplated in this study it is possible to conclude that they are different, that is, they were not generated by the same stochastic process.
38
Sato, Joao Ricardo. "Processos com memória longa compartilhada." Universidade de São Paulo, 2004. http://www.teses.usp.br/teses/disponiveis/45/45133/tde-22082007-160332/.
Full textAbstract:
Este trabalho tem como objetivo a avaliação de três estimadores do parâmetro de integração fracionária d e de um teste para memória longa compartilhada. Os estimadores a serem avaliados são: o estimador de Geweke e Porter-Hudak, o estimador usando o periodograma suavizado e o estimador semiparamétrico truncado de Whittle. A avaliação dos estimadores será no contexto de processos ARFIMA+ARMA, e em relação a variações nos termos autoregressivos e de médias móveis, tanto do termo de memória curta quanto do termo de memória longa. Além disso, serão introduzidos o conceito de modelos com memória longa compartilhada e um método de identificação através da análise de correlação canônica para séries temporais multivariadas proposto, por Ray e Tsay (1997). Por fim, serão apresentadas três aplicações sobre dados reais dos tópicos estudados: uma para a velocidade do vento em São Paulo e Piracicaba e outras duas para séries das bolsas de valores de Hong Kong, Nova Zelândia, Singapura, Brasil e Reino UnidoThe goal of this project is the evaluation of three long memory parameter estimators and a common long range dependence test. The estimators evaluated are: the Geweke and Porter-Hudak, the smoothed periodogram and the semiparametric truncated Whittle estimators. The evaluation is in the context of processes ARFIMA+ARMA, and related to variations in the autoregressive and moving average coefficients, both in the short and long memory terms. Furthermore, we describe common long range dependence processes and an identification approach (Ray and Tsay, 1997) for them, using the canonical correlation analysis. Finally, three applications to real data are presented: the first one to the wind\'s speed in the Brazilian cities of São Paulo and Piracicaba, and the other ones to financial time series of the stock markets of Hong Kong, New Zealand, Singapore, Brazil and the United Kingdom.
39
Chmelík, Pavel. "Mají odkupy zbraní pozitivní vliv na míru kriminality?" Master's thesis, Vysoká škola ekonomická v Praze, 2013. http://www.nusl.cz/ntk/nusl-199725.
Full textAbstract:
This paper analyzes effect of gun buyback that took place in Great Britain in years 1996 and 1997 on crime rate and compares the results with theoretical arguments and previous empirical findings. It contains analysis of three independent time series: crime rate in England and Wales, Scotland and Northern Ireland. Models of the time series are built using Box-Jenkins methodology. The models are tested for presence of a structural break using visual analysis, Chow test and Quandt-Andrews test. These tests are used as an evaluation criterion of the effect of buyback on crime rate. The result of the analysis is that it is not possible to reject the null hypothesis that buybacks do not have effect on crime rate.
40
Cunha, Manuel Maria Barbas Gaio Cardote da. "Projeções de migrações em Portugal e determinantes da sua evolução." Master's thesis, Instituto Superior de Economia e Gestão, 2020. http://hdl.handle.net/10400.5/20747.
Full textAbstract:
Mestrado em Métodos Quantitativos para a Decisão Económica e EmpresarialAtualmente, os fluxos migratórios são o fator que determina a dinâmica da população em Portugal, sendo, contudo, o processo demográfico mais volátil. A previsão de fluxos migratórios é uma tarefa multidimensional, caracterizada por elevados níveis de erro, não existindo uma teoria robusta para diferentes fluxos. O presente relatório de estágio tem como objetivo obter projeções para Portugal, entre 2018 e 2038. Para tal, foi aplicado o modelo proposto por Rogers e Castro (1981), bem como uma regressão linear com o intuito de identificar possíveis fatores explicativos. O cálculo dos parâmetros do modelo permite uma representação dos dados e uma melhor interpretação dos mesmos, sendo a sua projeção feita recorrendo a modelos auto-regressivos integrados de médias móveis. Os resultados alcançados apresentam conclusões positivas para o panorama atual da estrutura etária portuguesa, verificando-se um saldo migratório positivo em alguns dos cenários definidos. Apesar das limitações verificadas, foi também possível aferir o impacto significativo do desemprego no saldo migratório.
Nowadays, migratory flows are the factor that determines the population dynamics in Portugal, being, however, the most volatile demographic process. The prediction of migratory flows is a multidimensional task, characterized by high levels of error, and there is no robust theory for different flows. This internship report aims to obtain projections for Portugal, between 2018 and 2038. To this end, the model proposed by Rogers and Castro (1981) was applied, as well as a linear regression in order to identify possible explanatory factors. The calculation of the model parameters allows a representation of the data and a better interpretation of the same, being projected using autoregressive integrated moving average models. The results achieved show positive conclusions for the current panorama of the Portuguese age structure, with a positive net migration in some of the defined scenarios. Despite the verified limitations, it was also possible to assess the significant impact of unemployment on the net migration.
info:eu-repo/semantics/publishedVersion
41
Логін, Вадим Вікторович. "Моделі для прогнозування характеристик трафіка цифрової реклами." Master's thesis, Київ, 2018. https://ela.kpi.ua/handle/123456789/23748.
Full textAbstract:
Магістерська дисертація: 112 с., 48 рис., 40 табл., 3 додатки і 30 джерел.
Об’єкт дослідження – трафік цифрової реклами у формі статистичних даних.
Предмет дослідження – моделі та методи аналізу даних у формі часових рядів, методи прикладної статистики.
Мета роботи – побудова моделей часових рядів для прогнозування найважливіших характеристик трафіка цифрової реклами.
Методи дослідження – моделі часових рядів для прогнозування даних та порівняльний аналіз отриманих моделей.
У даній роботі наведені результати побудови моделей часових рядів, що призначені для прогнозування найважливіших характеристик трафіка цифрової реклами. Описані результати порівняльного аналізу отриманих моделей за допомогою інформаційних критеріїв, а також з точки зору їхньої точності. Встановлено, що для нашої задачі, найкращою моделлю є модель ARIMAX (Autoregressive integrated moving-average model with exogenous inputs), тобто модель авторегресії та ковзного середнього з екзогенними змінними. Тому для подальших досліджень рекомендовано використовувати саме цю модель.
За матеріалами магістерської дисертації були написані тези, а також написана наукова стаття. Тези будуть опубліковані в збірці тез доповідей конференції САІТ-2018. А наукова стаття буде опублікована в електронній збірці доповідей у видавництві CEUR.
Прогнозні припущення щодо подальшого розвитку об’єкта дослідження – побудова нових, а також вдосконалення існуючих моделей часових рядів для прогнозування найважливіших характеристик цифрової реклами. А також узагальнення дослідження, що проводилось у даній роботі, на аналіз окремих сайтів із рекламного трафіку.Models for forecasting parameters of digital advertising traffic. Master's thesis: 112 p., 48 fig., 40 tabl., 3 appendixes and 30 sources. The object of study – digital advertising traffic in the form of statistical data. Subject of research – models and methods of analysis of data in the form of time series, methods of applied statistics. Purpose – constructing time series models for forecasting the most important characteristics of digital advertising traffic. Methods of research – time series models for forecasting data and comparative analysis of the obtained models. This paper presents the results of construction of time series models, which are intended for forecasting of the most important characteristics of digital advertising traffic. Described the results of the comparative analysis of the obtained models with the help of information criteria, and also in terms of their accuracy. Was found that for our task, the best model is the ARIMAX model (Autoregressive integrated moving-average model with exogenous inputs). Therefore, it is recommended to use this model for further research. Based on master's dissertation were written theses as well as a scientific article. The theses will be published in the SAIT-2018 conference Book of Abstracts. The scientific article will be published in the electronic collection of reports at the CEUR publishing house (CEUR Workshop Proceedings). The further development of the research object – is the construction of new ones, as well as the improvement of existing time series models for forecasting the most important characteristics of digital advertising traffic. And also – it is a generalization of the research, conducted in this paper, on the analysis of individual sites from the digital advertising traffic.
42
Fabris, Thiago Rocha. "A projeção dos lucros trimestrais para as companhias brasileiras através de modelos ARIMA." Florianópolis, SC, 2009. http://repositorio.ufsc.br/xmlui/handle/123456789/92675.
Full textAbstract:
Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Sócio-Econômico, Programa de Pós-Graduação em Economia, Florianópolis, 2009.Made available in DSpace on 2012-10-24T11:04:58Z (GMT). No. of bitstreams: 1 266095.pdf: 627717 bytes, checksum: 86fa9f39c4971b1fdf99b307f68a5022 (MD5)
O estudo trata da aplicação da metodologia Box e Jenkins (1970) para a previsão das séries dos lucros em companhias de capital aberto no Brasil. Diversos autores como Watts (1975), Foster (1977), Griffin (1977) e Brown e Rozeff (1979) têm sugerido que os lucros trimestrais podem ser previstos através de modelos ARIMA. Benston e Watts (1978) corroboraram em favor do modelo de Foster (1977) descrito por um SARIMA (100) x (010). Lorek (1979) argumenta em favor do modelo SARIMA (011) x (011) propostos por Griffin(1977) e Watts (1975). Collins e Hopwwod (1980) e Bathke e Lorek (1984) demonstraram que os lucros podem ser modelados por um processo SARIMA (100) x (011) sugerido por Brown e Rozeff (1979). As evidências empíricas recentes demonstram que os modelos random walk with drift (RWD) descritos por um ARIMA (100) e o processo SARIMA (011) x (011) dominam conjuntamente os outros modelos propostos na literatura pertinente, Lorek e Willinger (2007). O objetivo do trabalho foi averiguar se existe um modelo predeterminado que possa descrever o comportamento das séries temporais do lucro líquido e lucro operacional para determinados setores econômicos das empresas brasileiras. Concomitantemente analisar se existe um modelo padrão diferente dos modelos referidos acima. Os resultados encontrados corroboram com os modelos RWD para as séries do lucro líquido e Brown e Rozeff (1979) para a série do lucro operacional. Porém, a maior concentração dos modelos, via metodologia Box e Jenkins, podem ser descritos por processo ARIMA (000) x (000) e (010) x (000) para o lucro líquido e lucro operacional, respectivamente.
The study it treats on the application of the methodology Box and Jenkins (1970) for the forecast of the taxes of growth in company of capital opened in Brazil. Several authors as Watts (1975), Foster (1977), Griffin (1977) and Brown and Rozeff (1979) have suggested that the quarterly earnings can be be provided by ARIMA models. Benston and Watts (1978) corroborate the model in favor of Foster (1977) described a SARIMA (100) x (010). Lorek (1979) argues in favor of the model SARIMA (011) x (011) proposed by Griffin (1977) and Watts (1975). Collins and Hopwwod (1980) and Bathke and Lorek (1984) showed that profits can be modeled by a process SARIMA (100) x (011) suggested by Brown and Rozeff (1979). The recent empirical evidence shows that the random walk models with drift (RWD) described by an ARIMA (100) and a process SARIMA (011) x (011) jointly dominate the other models proposed in pertinent literature, Lorek and Willinger (2007). The objective of this work was check out if exists a predetermined model that can describe the behavior of the time series of the net earnings and operational earnings for definitive economic sectors of the Brazilian companies. Concomitantly to analyze if exists a standard different from the models mentioned above. The find results corroborate with models RWD, for the series of net earnings and Brown and Rozeff (1979) for the series of the operational earnings. However, the greatest concentration of the models, via Box and Jenkins methodology, can be described by ARIMA process (000) x (000) and (010) x (000) for the net earnings and operational earnings respectively.
43
Palandi, Victor Camillo. "Análise e projeção do ecommerce em Portugal." Master's thesis, Instituto Superior de Economia e Gestão, 2021. http://hdl.handle.net/10400.5/22752.
Full textAbstract:
Mestrado Bolonha em Métodos Quantitativos para a Decisão Económica e EmpresarialO consumo online é pauta relevante na sociedade desde o início dos anos 2000. Potencializado pela pandemia global, a importância estratégica deste canal para todos os agentes de mercado é indiscutível. O projeto de pesquisa tem como objetivo apresentar a realidade e evolução do e-commerce em Portugal, a partir da análise de um painel de domicílios, bem como prever a evolução de vendas do canal em 2021. São aplicadas metodologias de alisamento exponencial e modelos de previsão ARIMA de Box-Jenkins a uma base de painel de domicílios concedida pela NielsenIQ - líder mundial em pesquisa de mercado. Conforme espetável, o estudo aponta para uma curva ascendente a nível de vendas do canal até o final de 2021 e deve ser alvo determinante para uma estratégia de sucesso de retalhistas e indústria, bem como uma necessidade latente por parte do consumidor.
Online consumption has been a relevant issue in society since the early 2000s. Powered by the global pandemic, the strategic importance of this channel for all market agents is remarkable. This project aims to present the reality and evolution of e-commerce in Portugal, from the analysis of a panel of households, as well as to predict the evolution of the channel's sales in 2021. Exponential smoothing methodologies and models of Box- Jenkins ARIMA are applied in a household panel database provided by NielsenIQ - world leader in market research. As expected, the study points to an upward curve in the channel's sales by the end of 2021 and should be a key target for a successful strategy for retailers and industry, as well as a latent demand for the consumer.
info:eu-repo/semantics/publishedVersion
44
Conway, Eunan Martin. "Stochastic modelling and forecasting of solar radiation." Thesis, Northumbria University, 1998. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.367414.
Full text
45
Kern, Joshua Victor. "The Development of Measurement and Characterization Techniques of Road Profiles." Thesis, Virginia Tech, 2004. http://hdl.handle.net/10919/10118.
Full textAbstract:
The principal excitation to a vehicle's chassis system is the road profile. Simulating a vehicle traversing long roads is impractical and a method to produce short roads with given characteristics must be developed. By understanding the characteristics of the road, a reduced set of models can be created from which appropriate representations of the terrain can be synthesized. Understanding the characteristics of the terrain requires the ability to accurately measure the terrain topology. It is only by increasing the fidelity and resolution of terrain topology data that application of these data can be advanced. The first part of this work presents the development of a high fidelity 3-D laser terrain measurement system. The system is developed for both on-highway and off-road measurement. It is capable of measuring terrain in three dimensions, whereas current systems measure separate 2-D profiles in each wheel path of the vehicle. The equipment setup and signal processing techniques are discussed, as well as future improvements and applications of this enabling technology.
The second part of this work develops a method of characterizing non-stationary road profile data using ARIMA (Autoregressive Integrated Moving Average) modeling techniques. The first step is to consider the road to be a realization of an underlying stochastic process. The model identification techniques are demonstrated. Statistical techniques are developed and used to examine the distribution of the residual process and the results are demonstrated. The use of the ARIMA model parameters and residual distributions in classifying road profiles is also discussed. By classifying various road profiles according to given model parameters, any synthetic road realized from a given class of model parameters will represent all roads in that set, resulting in a timely and efficient simulation of a vehicle traversing any given type of road.Master of Science
46
Sá, Vânia Catarina Neves de. "O desemprego jovem em Portugal." Master's thesis, FEUC, 2014. http://hdl.handle.net/10316/25429.
Full textAbstract:
Trabalho de projeto de mestrado em Economia (Economia Financeira), apresentado à Faculdade de Economia da Universidade de Coimbra, sob a orientação de António Manuel Portugal Duarte.O presente trabalho aborda uma temática bastante atual que se assume como uma das maiores preocupações dos decisores políticos e da população portuguesa: o desemprego e, em particular, o desemprego jovem. O problema do desemprego jovem atinge não só Portugal, mas também vários países da União Europeia, em que o número de desempregados jovens tem vindo a aumentar de forma acentuada. Com este estudo pretendemos dar a conhecer a realidade que nos rodeia através da comparação e análise de dados estatísticos relativos à taxa de desemprego jovem, bem como averiguar as possíveis causas e consequências do fenómeno. Desta forma foi possível verificar que o desemprego jovem tem apresentado um crescimento algo exponencial, sobretudo a partir de 2008, ano em que Portugal foi atingido pela atual crise económica e financeira, registando quase 40% de jovens desempregados, em 2012. Esta situação parece assim refletir a estagnação económica que o nosso país está a atravessar, sendo que o propósito do nosso estudo passa também por apresentar algumas medidas de política económica tendo em vista a diminuição deste flagelo que provoca consequências muito importantes a diversos níveis, económico, político e social.
47
Hsieh, shu ju, and 謝淑如. "Application and Integration of Consecutive ARIMA Transfer and Seasonal ARIMA Trnasfer Function." Thesis, 1993. http://ndltd.ncl.edu.tw/handle/00723108764295798857.
Full text
48
楊耀華. "ARIMA-BP times series neural networks." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/31461568317667656927.
Full textAbstract:
碩士中華大學
資訊管理學系(所)
96
In this paper we proposed an ARIMA-BPN algorithm combining advantages of ARIMA and Back-propagation networks (BPN). The algorithm is based on BPN and its inputs are the same as ARIMA. It can generate a non-linear function to create an accurate model to predict time series. The BPN algorithm must be modified because the residuals would be changed when the weights were changed during continuously training BPN. That is we will use the continuously updated residuals as inputs. This study examined 6 artificial designed cases and 4 real world cases to evaluate the abilities of the ARIMA, BPN, and ARIMA-BPN. The results sowed that ARIMA-BPN is the most accurate methods in some cases.
49
Mendes, João Filipe Batista. "Forecasting bitcoin prices: ARIMA vs LSTM." Master's thesis, 2019. http://hdl.handle.net/10071/19724.
Full textAbstract:
Bitcoin has recently received special attention in economics and finance as the most popular blockchain technology. This dissertation aims to discuss whether newly machine-leaning models perform better than traditional models in forecasting. Particularly, this study compares the accuracy of the prediction of bitcoin prices using two different models: Long-Short Term Memory (LSTM) versus Auto Regressive Integrated Moving Average (ARIMA), in terms of forecasting errors, and Python routines were used for such purpose. Bitcoin price time series ranges from 2017-06-18 to 2019-08-07, in a daily basis, sourced from the Federal Reserve Economic Data. To compare the results of both models, data was divided into two subsets: training (83.5%) and testing (16.5%). The literature usually indicates that LSTM outperforms ARIMA. In this dissertation, the results do confirm that LSTM forecasts of bitcoin prices improve on average ARIMA predictions by 92% and 94%, according to RMSE and MAE.A Bitcoin tem recebido recentemente especial atenção em áreas como a economia e finanças por ser a mais popular tecnologia de blockchain. Esta dissertação tem como objetivo verificar se os novos modelos de machine-learning apresentam melhores resultados que os modelos tradicionais em previsões. Este estudo compara, em particular, a precisão da previsão do preço da Bitcoin usando dois modelos diferentes: Long-Short Term Memory (LSTM) versus Auto Regressive Integrated Moving Average (ARIMA), em termos de erros de previsão e aplicando rotinas do Python. A análise teve como base os preços diários da Bitcoin entre 18 de junho de 2016 e 7 de agosto de 2019, retirados da base de dados da Reserva Federal. Para comparar os resultados dos dois modelos, os dados foram divididos em duas secções: o treino (83.5%) e o teste (16.5%). A literatura indica que o modelo LSTM tem uma melhor precisão que o ARIMA e nesta dissertação os resultados confirmam que o modelo LSTM melhora em média 92% e 94% a previsão do ARIMA, de acordo com o RMSE e o MAE.
50
Tseng, Shuhui, and 曾淑惠. "Structural Change ARIMA Modeling and Application." Thesis, 1993. http://ndltd.ncl.edu.tw/handle/44658612423888170873.
Full textAbstract:
碩士國立政治大學
統計學研究所
81
Non-linear time series analysis is a rapidly developing subject in recent years. One of special families of non-linear models is threshold model. Many literatures have shown that even simple threshold model can describe certain types of time series, such as structural change behavior, more faithful than using linear ARMA models. In this paper, we discuss some problems about the threshold model and structural change analysis. Instead of finding the change point, we present the change period concepts on the model- building. An efficient algorithem on constructing the structure change ARIMA models is proposed. Finally, we demonstrate an example about the birth rate of Taiwan, and the comparison of forecasting performance for the structure change ARIMA model with alternative models are also made.