Thèses sur le sujet « Indici similarità »

L'elaborato tratta l'applicazione del metodo di diagnostica predittiva model of signal based ai dati delle vibrazioni di una macchina automatica. La vibrazioni della macchina sono state rappresentate per mezzo di modelli autoregressivi (AR) e per la stima dei parametri è stato utilizzato il metodo dei minimi quadrati. Per il confronto tra i modelli della macchina in funzionamento nominale e quelli in condizioni anomale, sono stati utilizzati vari indici di similarità, tra i quali, gli indici spettrali che hanno fornito maggiorni informazioni rispetto ai restanti. Tale procedimento è stato implementato anche su PLC Beckhoff in linguaggio ST.

L’attuale gold standard diagnostico nella routine clinica utilizzato per escludere la natura maligna di noduli tiroidei, è rappresentato dalla valutazione morfologica del materiale ottenuto da biopsie. Tuttavia, non sempre è possibile arrivare ad una diagnosi citologica affidabile e circa il 20-30% dei noduli risultano “inderterminati per malignità”. I pazienti con questa diagnosi vengono quindi sottoposti a tiroidectomia totale e dopo l’analisi istologica post-operatoria l’80% circa di essi risultano essere benigni. L’impatto dell’operazione sul paziente è rilevante poiché le funzioni fisiologiche della tiroide dovranno essere sostituite cronicamente con l’utilizzo di farmaci, il cui costo, in aggiunta a quello dell’operazione, ha un peso sul bilancio sanitario. Negli ultimi anni, nel campo della ricerca biomedica, grande attenzione è stata riposta verso l’analisi proteomica e verso la sua potenziale applicazione nella ricerca di nuovi biomarcatori, determinanti nel discriminare noduli benigni da maligni in modo da minimizzare la diagnosi di malignità indeterminata. La spettrometria di massa è uno degli strumenti più importanti per ottenere informazioni riguardanti la composizione molecolare di un campione, la presenza di biomolecole e la loro abbondanza. Tra i diversi approcci proteomici in grado di identificare alterazioni molecolari di diversi tipi di lesioni, la tecnica di Imaging MALDI-MSI (Matrix Assisted Laser Desorption/Ionisation-Mass Spectrometry Imaging) ha guadagnato sempre più spazio e interesse. MALDI-MSI rappresenta una tecnologia ideale che permette di esplorare la distribuzione spaziale di biomolecole nel tessuto integrando informazioni molecolari e quelle tradizionali morfologiche. Visti i recenti risultati ottenuti tramite l’analisi MALDI-MSI di campioni di tessuto tiroideo nell’identificazione di segnali proteomici in grado di discriminare casi benigni da maligni, l’idea che è nata è stata quella di applicare per la prima volta questo tipo di analisi a campioni citologici ottenuti da biopsie di noduli tiroidei. Prima di poter applicare la tecnica MALDI allo studio clinico, il protocollo di analisi è stato ottimizzato per evitare problemi di degradazione, fenomeni di alterazioni o contaminazioni e formazione di artefatti. Due diversi problemi tecnici quali i) l’interferenza dell’emoglobina a causa dell’elevata vascolarizzazione dell’organo e ii) la stabilità del campione nel tempo prima dell’analisi da un punto di vista morfologico e proteomico, sono stati affrontati e risolti in due studi pianificati come parte del progetto di tesi. In origine lo studio clinico per l’identificazione di potenziali cluster di segnali con proprietà discriminanti prendeva in considerazione un ampio numero di campioni di noduli tiroidei ma, a causa del lento arruolamento di casi maligni per la loro natura rara, la tesi contiene solo i risultati di un analisi preliminare. 18 soggetti sono stati arruolati per il training set (9 noduli benigni e 9 maligni). Il modello di regressione logistica penalizzato (LASSO) è stato costruito su un set di dati di 81 regioni di interesse, in accordo con l’identificazione morfologica operata dal patologo per evitare false informazioni derivanti da cellule diverse dai tirociti. Il modello di classificazione è stato validato su 11 pazienti con diversi tipi di lesioni (i.e. benigna, indeterminata e maligna). I risultati sono molto promettenti e sottolineano la possibilità di introdurre MALDI-MSI come uno strumento complementare nella caratterizzazione diagnostica delle lesioni tiroidee. Sono inoltre stati esaminati gli indici di similarità più utilizzati tra i numerosi profili di spettri ed è stata proposta una nuova misura. Uno studio di simulazione di spettri di massa è stato poi implementato per identificare le migliori misure di similarità in termini di performance, da applicare per comparare profili proteomici.
The actual gold standard to exclude the malignant nature of thyroid nodules in the clinical routine is represented by thyroid Fine Needle Aspirations (FNAs) biopsies. Thyroid FNAs are safe and cost-effective. Approximately the 20-30% of cases have an indeterminate for malignancy final report. These patients undergo diagnostic (and not therapeutic) thyroidectomy, but after surgery the 80% of these thyroid nodules are benign. This overtreatment has of course important consequences in the quality of life of the patients and high healthcare costs. The application of -omics techniques might have a potential role in the research for new diagnostic markers able to discriminate benign from malignant nodules, thus minimizing the challenging cases of indeterminate for malignancy. Mass spectrometry is one of the most important analytical tools able to obtain information regarding the molecular composition of a sample, the presence of biomolecules and their abundance. Among the different proteomics approaches able to extract the molecular alterations of the different type of specimen’s lesion, Matrix-Assisted Laser Desorption/Ionization (MALDI) Mass Spectrometry Imaging (MSI) was strongly emerging. MALDI-MSI represents an ideal technology that enables to explore the spatial distribution of biomolecules within tissue, integrating molecular and traditional morphological information while preserving the integrity of the analysed tissue. Various studies applied MALDI-MSI technology for prognostic purposes and for in real time diagnostic setting, showing the usefulness, advantages and applicability of MALDI-MSI in different fields of pathology. Due to the promising results recently obtained with MALDI-MSI in the identification of proteomic signals able to differentiate between benign and malignant cases from the analysis of thyroid tissue after surgery , the idea was to apply for the first time MALDI-MSI on real thyroid FNAs biopsies. Preliminary to the clinical study, the protocol for the proteomic MALDI-MSI analysis was optimised to avoid degradation, alteration phenomena, contamination and artefacts formation. The methodological improvement of the protocol in a complicated field as thyroid cytological specimens played an important role in this study. Challenging technical aspects, such as i) the interference of haemoglobin due to the high vascularization of the thyroid organ and ii) the stability of the samples over time before the analysis from a morphological and proteomic point of view, were overcome through two studies that were planned and analysed as part of the thesis. The clinical study for the detection of the potential cluster of signals with discriminant capability was originally planned to involve a large sample of thyroid nodules, however, due to the slow enrolment rate of malignant cases, the thesis contains only the results of a preliminary analysis. Eighteen subjects contributed to the training set with 9 benign and 9 malignant thyroid nodules. The statistical model was based on data of 81 specific region of interest, according to the morphological triage performed by the pathologist in order to overcome false information deriving from non thyrocytes cells. The validation phase was performed on 11 patients with different type of lesions (i.e. benign, indeterminate and malignant). Results are very promising and highlight the possibility to introduce MALDI-MSI as a complementary tool for the diagnostic characterization of thyroid lesions. A methodological aspect that emerged from the peculiarity of the proteomic analysis was also investigated. A review of the most used indices for the assessment of the similarity between mass spectra profiles was performed and a new measure was proposed. A simulation study was implemented in order to identify the best similarity measure to use in comparing proteomic profiles.

Il presente elaborato si propone di fornire un’esaustiva trattazione relativamente a quella che viene chiamata produzione cellulare, o Cellular Manufacturing, conosciuta come un’importante applicazione della Group Technology. Ne vengono discusse le condizioni di applicazione, i conseguenti benefici apportati e le molteplici metodologie, presentate in letteratura, per affrontare il problema della formazione delle celle produttive (Cell Formation Problem). L’obiettivo dell’elaborato, però, non si limita solamente a descrivere i vari aspetti citati, in quanto si è tentato di automatizzare, con successo, tali procedure attraverso la creazione della piattaforma software, New Cluster Calculator, sviluppata all’interno dell’ambiente di sviluppo integrato (IDE) di Microsoft Visual Studio, il cui scopo è proprio quello di proporre soluzioni, in tempi ragionevoli, al problema della formazione cellulare. L’approccio adottato dal programma consiste nell’impiego di alcuni coefficienti di similarità, utili a valutare la somiglianza delle macchine da raggruppare, in combinazione con i principali algoritmi di clustering. Per l’assegnamento delle famiglie di prodotti alle celle, invece, si è utilizzata una procedura euristica. Infine, si sono realizzate varie simulazioni al fine di poter effettuare valutazioni inerenti alla qualità delle soluzioni ottenute e dimostrare la scalabilità del software, in quanto, anche per istanze che considerano un numero importante di input, New Cluster Calculator risulta uno strumento molto efficiente. Inoltre, si è giunti alla conclusione che, poiché le fasi analizzate durante l’elaborato appartengono ad un processo gerarchico, i risultati, in termini di prestazione, dipendono dall’efficacia dei passaggi a monte e dalla loro corretta integrazione. In altre parole, il frutto di ogni step decisionale dipende dalla qualità dei suoi dati di input.

Cette thèse questionne les manifestations de la convergence (i.e. le rapprochement entre les productions des participants) au niveau interactionnel. Pour cela, les termes d'alignement (défini en rapport avec l'activité en cours) et d'affiliation (l'expression d'un même stance par les participants) sont empruntés à l'Analyse Conversationnelle. Le corpus utilisé est le CID-Corpus of Interational Data, corpus de conversation (interaction non-contrainte, hautement coopérative et globalement symétrique).Nous interrogeons le lien entre la convergence et la similarité lexicale, grâce à l'analyse d'une collection de 300 hétéro-répétitions (recueillie grâce à un outil d'aide au repérage des répétitions). Nous proposons ensuite une analyse quantitative de l'évolution des réponses des auditeurs, puis une analyse qualitative de discours rapportés directs, phénomènes susceptibles de faire émerger de l'affiliation. Nous montrons que les hétéro-répétitions lexicales et les discours rapportés « en écho » (discours rapportés produits par l'auditeur de la narration) peuvent être utilisés (entre autres) pour exprimer l'alignement et l'affiliation, ce qui, en cas de ratification, crée les conditions propices à l'émergence d'un moment de convergence interactionnelle. Nous montrons également que ces mêmes phénomènes peuvent servir à créer le désalignement temporaire nécessaire à l'engagement dans une séquence oblique convergente. Ainsi, ce travail décrit l'établissement et le fonctionnement de séquences convergentes, à travers l'étude de phénomènes interactionnels méconnus
This thesis investigates the manifestations of convergence (i.e. the rapprochement between the participants' productions) at the level of interaction. With this aim, the terms of alignment (defined in relation to the current activity) and affiliation (display of the same stance by both participants) are borrowed from Conversation Analysis. The conversational corpus (non-constrained, highly cooperative and globally symmetrical interaction) used is the CID-Corpus of Interactional Data. Firstly, the link between convergence and lexical similarity is investigated thanks to the analysis of a collection of 300 other-repetitions (collected using a tool to assist in the detection of OR). Secondly, storytelling is studied and a quantitative analysis of the evolution of listeners' responses is proposed together with a qualitative analysis of direct reported speech phenomena, which are likely to make affiliation emerge. These analyses show that lexical other-repetitions and "echo" reported speech (reported speech which is produced by the listener of the narrative) can be used by participants to, inter alia, express alignment and affiliation, which, in case of ratification, creates the adequate conditions for the emergence of interactional convergence. The same phenomena can be used to create the temporary disalignment necessary to engage in an oblique (and potentially convergent) sequence. This work then describes the establishment and the conduct of convergent sequences through the analysis of interactional phenomena

A study of the phytoplankton communities of Lake Ray Roberts was conducted at three sampling sites on two sampling dates during the summer of 1995, assessing both community structure and chlorophyll a concentrations. Eight similarity indices were then used to characterize and compare the communities of these sites. Both community similarity and chlorophyll a concentrations were evaluated using a minimum detectable difference equation to determine the amount of change that must occur in these parameters in order to be deemed statistically significant. The Bray-Curtis Index was shown to be the most adequate index evaluated, and was subsequently used in conjunction with bootstrap analysis to determine the similarity between the three sampling sites.

Malgré de nombreux travaux, il apparaît que les outils d'Analyse Typologiques ne permettent pas de réduire les délais d'implantation de la Technologie de Groupe. Cette démarche de rationalisation du système de production pose un problème général de classification que nous baptiserons ACCI. Il s'agit de construire sur un domaine de connaissance peu formalisé, une classification jugée pertinente par les experts du domaine. En terme d'Acquisition de Connaissances, l'ACCI a pour objectif de représenter l'activité d'un groupe d'expert à l'aide d'un modèle dit de classification simple. Les outils susceptibles de contribuer à résoudre ce problème proviennent de domaines variés : l'Analyse Typologique, la Classification Conceptuelle, les réseaux de Neurones et les Algorithmes Génétiques mais reposent sur un principe commun qui défini le concept d'Outils de Classification Automatique (OCA). Dans le cadre de l'ACCI, ces outils présentent des limites intrinsèques. La plupart des travaux menés pour y remédier, proposent des approchent interactives qui aident les experts à reformuler leurs connaissances en fonction de l'OCA utilisé. C'est à dire, sous la forme d'attributs descriptifs des objets tels qu'il devient possible d'en déduire des classes par des opérations de comparaison simples. Nous proposons une approche interactive qui a pour vocation de s'adapter aux experts en intégrant directement dans le processus de classification, les connaissances qu'ils sont à même de formuler naturellement lorsqu'ils valident une partition des objets du domaine. Cette approche a été testé sur un cas réel à l'aide d'un prototype informatique. Elle contribue à résoudre le problème de l'utilisation des outils d'Analyse Typologique pour réduire les délais d'implantation de la TG, et s'applique potentiellement à d'autre domaine, car les connaissances manipulées par le système ne sont pas liées au domaine d'application, mais à la structure recherchée (une partition).

Le travail de thèse qui est rapporté dans le présent document a porté sur le problème de l'individualisation des HRTF pour la synthèse binaurale. Les HRTF sont les filtres linéaires, chacun associé à une direction de l'espace, qui portent en eux l'expression de tous les indices physiques de localisation nécessaires pour une perception de l'espace par le système auditif. La synthèse binaurale utilise avantageusement ces filtres pour sculpter les signaux à présenter aux tympans de l'auditeur, afin de lui procurer l'illusion d'une scène sonore réaliste. Les HRTF étant très liées à la morphologie de la tête et des pavillons, la spatialisation n'est correctement assurée que si ces filtres sont bien adaptés à l'auditeur. Cependant, la mesure exhaustive des HRTF est coûteuse et inconfortable, et il s'agit donc de développer des moyens alternatifs pour les obtenir : c'est le problème de l'individualisation. On se focalise sur les indices spectraux de la localisation auditive, c'est-à-dire les colorations du spectre à dépendance directionnelle, qui constituent la part des HRTF la plus complexe et la plus variable d'un individu à l'autre. Le constat fondateur de nos investigations est le suivant: bien que les HRTF présentent des caractéristiques intrinsèquement individuelles, on peut dégager des évolutions spatiofréquentielles de leur spectre d'amplitude, communes d'un individu à l'autre, mais susceptibles d'être masquées par deux sources importantes de variabilité, que sont la taille et l'orientation des pavillons. Nous proposons des outils permettant de dépasser ces différences apparentes, afin de se focaliser sur ce qui est vraiment spécifique à chaque individu. Deux solutions techniques d'individualisation des HRTF sont développées en utilisant avantageusement la diversité des comportements offerte par les HRTF d'une base de données. La première solution proposée permet d'adapter, pour un nouvel auditeur, les HRTF d'un autre individu issues d'une base de données, en leur appliquant des transformations guidées par une comparaison morphologique entre les pavillons des deux sujets. Les hypothèses de travail et les outils proposés pour mettre en oeuvre la technique sont validés objectivement grâce aux données recueillies sur 6 sujets, et on montre que la méthode d'adaptation proposée dépasse les performances de l'état de l'art. La seconde solution permet de reconstruire les HRTF d'un nouvel auditeur pour une direction quelconque de l'espace à partir d'un nombre réduit de HRTF individuelles mesurées. La technique proposée est basée sur une base de données constituée des HRTF mesurées finement sur une centaine de sujets, à partir desquelles on génère des prototypes. La reconstruction des HRTF repose sur un processus de reconnaissance de formes entre les HRTF individuelles mesurées et ces prototypes. Une validation objective montre que, selon différents critères, les performances de reconstruction de la technique proposée dépassent celles de l'état de l'art. Ces résultats sont confirmés par une évaluation subjective, menée selon un protocole novateur en synthèse binaurale dynamique
This Ph. D. Thesis deals with the problem of Head-Related Transfer Functions (HRTFs) individualization, in the context of binaural synthesis. HRTFs embed ail the acoustical phenomena occurring on the path between a source at a given position in space and the listener's eardrums. As these linear filters convey all free field localization cues needed by the auditory system to perceive a 3D sound scene, HRTF can be used to sculpt the signals to be reproduced over headphones in order to create convincing spatialized auditory displays : this is the aim of binaural synthesis. HRTFs strongly depend on idiosyncratic morphological features (overall shape of the head, fine structure of the pinnae), and as a result, the use of non-individual HRTFs often leads to perceptual artifacts. Unfortunately, exhaustive acoustic measurements of individual HRTFs are long and uncomfortable for subjects, and it is therefore expected to develop alternative techniques to obtain customized HRTFs : this is the problem of individualization. As they represent the most complex and the most individual part of HRTFs, our study focusses on the colorations induced by pinna filtering, known as spectral cues. The founding assumption of our work is the following : although HRTFs contain intrinsically individual features, common spatio-frequential behaviours can be found from subject to subject. Such similarities may be hidden by the existence of two morphological sources of variability, being the size and orientation of ear pinnae. We develop tools whose aim is to go beyond apparent differences, and to focus on what is really specific of each individual. We propose two technical solutions for HRTF individualization, based on the use of a HRTF database. The first solution uses a 3D model-based morphological matching of pinnae shapes, to properly adapt existing non-individual HRTFs from a database, so that they fit to a new listener. To transform HRTF data, we propose a combination of frequency scaling and rotation shift, whose parameters are predicted by the result of the morphological comparison. The method is designed on the basis of data acquired from six subjects, and it is shown objectively that a better customization is achieved compared to the state-of-the-art technique. The second solut ion aims at reconstructing HRTF for any direction, from only sparse individual HRTF measurements. In order t o overcome the performance of classical blind interpolation techniques, additional knowledge is injected in the reconstruction process :HRTF prototypes are first extracted from the analysis of a large HRTF database, and serve as a well-informed background in a pattern recognition process. An objective assessment shows that , compared to previously developped techniques, HRTF reconstruction achieves a better spatial fidelity with the proposed method. FinaIly, this result is confirmed by a subjective evaluation based on a new protocol

Ce travail porte sur l'estimation des indices d'autosimilarité et de stabilité d'un processus ou champ stable fractionnaire et autosimilaire ou d'un processus stable multifractionnaire.Plus précisément, soit X un processus ou un champ stable H-autosimilaire à accroissements stationnaires (H-sssi) ou un processus stable multifractionnaire. Nous observons X aux points k/n, k=0,..., n.Nos estimations sont basées sur des variations de puissances négatives beta avec -1/2
This work is concerned with the estimation of the self-similarity and the stability indices of a H-self-similarity stable process (field) or a multifractional stable process.More precisely, let X be a H-sssi (self-similar stationary increments) symmetric alpha-stable process (field) or a multifractional stable process. We observe X at points k/n, k=0,...,n.Our estimates are based on beta-negative power variations with -1/2

La réussite du lancement d'un grand nombre des satellites Radar à Synthèse d'Ouverture (RSO - SAR) de nouvelle génération a fourni régulièrement des images SAR et SAR polarimétrique (PolSAR) multitemporelles à haute et très haute résolution spatiale sur de larges régions de la surface de la Terre. Le système SAR est approprié pour des tâches de surveillance continue ou il offre l'avantage d'être indépendant de l'éclairement solaire et de la couverture nuageuse. Avec des données multitemporelles, l'information spatiale et temporelle peut être exploitée simultanément pour rendre plus concise, l'extraction d'information à partir des données. La détection de changement de structures spécifiques dans un certain intervalle de temps nécessite un traitement complexe des données SAR et la présence du chatoiement (speckle) qui affecte la rétrodiffusion comme un bruit multiplicatif. Le but de cette thèse est de fournir une méthodologie pour simplifier l'analyse des données multitemporelles SAR. Cette méthodologie doit bénéficier des avantages d'acquisitions SAR répétitives et être capable de traiter différents types de données SAR (images SAR mono-, multi- composantes, etc.) pour diverses applications. Au cours de cette thèse, nous proposons tout d'abord une méthode générale basée sur une matrice d'information spatio-temporelle appelée Matrice de détection de changement (CDM). Cette matrice contient des informations de changements obtenus à partir de tests croisés de similarité sur des voisinages adaptatifs. La méthode proposée est ensuite exploitée pour réaliser trois tâches différentes: 1) la détection de changement multitemporel avec différents types de changements, ce qui permet la combinaison des cartes de changement entre des paires d'images pour améliorer la performance de résultat de détection de changement; 2) l'analyse de la dynamicité de changement de la zone observée, ce qui permet l'étude de l'évolution temporelle des objets d'intérêt; 3) le filtrage nonlocal temporel des séries temporelles d'images SAR/PolSAR, ce qui permet d'éviter le lissage des informations de changement dans des séries pendant le processus de filtrage.Afin d'illustrer la pertinence de la méthode proposée, la partie expérimentale de la thèse est effectuée sur deux sites d'étude: Chamonix Mont-Blanc, France et le volcan Merapi, Indonésie, avec différents types de changements (i.e. évolution saisonnière, glaciers, éruption volcanique, etc.). Les observations de ces sites d'étude sont acquises sur quatre séries temporelles d'images SAR monocomposantes et multicomposantes de moyenne à haute et très haute résolution: des séries temporelles d'images Sentinel-1, ALOS-PALSAR, RADARSAT-2 et TerraSAR-X
A large number of successfully launched and operated Synthetic Aperture Radar (SAR) satellites has regularly provided multitemporal SAR and polarimetric SAR (PolSAR) images with high and very high spatial resolution over immense areas of the Earth surface. SAR system is appropriate for monitoring tasks thanks to the advantage of operating in all-time and all-weather conditions. With multitemporal data, both spatial and temporal information can simultaneously be exploited to improve the results of researche works. Change detection of specific features within a certain time interval has to deal with a complex processing of SAR data and the so-called speckle which affects the backscattered signal as multiplicative noise.The aim of this thesis is to provide a methodology for simplifying the analysis of multitemporal SAR data. Such methodology can benefit from the advantages of repetitive SAR acquisitions and be able to process different kinds of SAR data (i.e. single, multipolarization SAR, etc.) for various applications. In this thesis, we first propose a general framework based on a spatio-temporal information matrix called emph{Change Detection Matrix} (CDM). This matrix contains temporal neighborhoods which are adaptive to changed and unchanged areas thanks to similarity cross tests. Then, the proposed method is used to perform three different tasks:1) multitemporal change detection with different kinds of changes, which allows the combination of multitemporal pair-wise change maps to improve the performance of change detection result;2) analysis of change dynamics in the observed area, which allows the investigation of temporal evolution of objects of interest;3) nonlocal temporal mean filtering of SAR/PolSAR image time series, which allows us to avoid smoothing change information in the time series during the filtering process.In order to illustrate the relevancy of the proposed method, the experimental works of the thesis is performed on four datasets over two test-sites: Chamonix Mont-Blanc, France and Merapi volcano, Indonesia, with different types of changes (i.e., seasonal evolution, glaciers, volcanic eruption, etc.). Observations of these test-sites are performed on four SAR images time series from single polarization to full polarization, from medium to high, very high spatial resolution: Sentinel-1, ALOS-PALSAR, RADARSAT-2 and TerraSAR-X time series

Nous avons cherché à enrichir l'Analyse Relationnelle (développée par Marcotorchino et Michaud) de différents outils (algèbre relationnelle, logique multivalente, analyse et optimisation combinatoire). Cette approche nous a permis d'étendre l'AR et ses champs d'application. Outre les applications classiques (classification automatique et agrégation multicritère), nous nous sommes intéressés à la théorie du choix social et à la théorie des tresses. Les contributions que nous présentons sont : Définition d'un nouvel algorithme de classification automatique; Unification de certains indices de similarité, et critères d'association entre variables qualitatives et ordonnées; Définition d'une nouvelle méthode d'agrégation, fondée sur une règle majoritaire et des approches combinatoires (Poincaré et Jordan); Généralisation des résultats de Black, d'Inada et de Sen en théorie du choix social; Modélisation originale des mouvements des brins d'une tresse et applications.

Made available in DSpace on 2016-06-02T19:31:54Z (GMT). No. of bitstreams: 1 3596.pdf: 1182246 bytes, checksum: 313badd826fcb755887e3f7becfbf99f (MD5) Previous issue date: 2011-03-04
Universidade Federal de Minas Gerais
Aiming to answer the question: preserved streams can sustain higher beta diversity than degraded streams? We estimated beta diversity of preserved and degraded streams during the rainy season of 2009 in pool and riffle areas using as substrate bags of leaf litter of tree species Talauma ovata. We analyzed the similarity in taxonomic composition of different mesohabitats using different indices of similarity (Sørensen, Morisita-Horn and Horn) and their corresponding estimators of diversity. Our study showed higher values of beta diversity in streams degraded when estimated by the similarity indices of Sørensen and Morisita-Horn. On the other hand the Horn index indicated no differences between the streams. This greater differentiation in faunistic composition between the different streams in degraded mesohabitats seems to be related to degradation of the riparian zone which favored the largest effect of hydrological disturbance due to increased volume of water especially in areas of pool resulting in elimination of taxa and increase local the dominance of more tolerant taxa, affecting beta diversity. This study paves the way for new understandings of the effects of degradation of riparian zones in streams, because the comparison between habitats within a given stream may provide a basis to assess their state of degradation, making possible new approaches to biomonitoring.
Com o objetivo de responder a pergunta: Córregos preservados podem sustentar uma maior diversidade beta que córregos degradados? Estimamos a diversidade beta de córregos preservados e degradados durante a estação chuvosa de 2009 em áreas de remanso e corredeira utilizando como substrato sacos de detritos foliares da espécie arbórea Talauma ovata. Analisamos a similaridade na composição taxonômica dos diferentes mesohabitats utilizando diferentes índices de similaridade (Sørensen, Morisita-Horn e Horn) e seus estimadores de diversidade correspondentes. Nosso estudo mostrou maiores valores de diversidade beta nos córregos degradados quando estimado pelos índices de similaridade de Sørensen e Morisita-Horn. Já o índice de Horn não indicou diferenças entre os córregos. Esta maior diferenciação na composição faunística entre os diferentes mesohabitats nos córregos degradados parece estar relacionada à degradação da zona ripária que favoreceu o maior efeito das perturbações hidrológicas devido ao maior volume de água especialmente nas áreas de remanso, resultando em eliminação local de táxons e aumento na dominância de táxons mais tolerantes, afetando a diversidade beta. Este estudo abre caminho para novas perspectivas de compreensão dos efeitos da degradação das zonas ripárias em córregos, pois a comparação entre habitats dentro de um dado córrego pode fornecer subsídios para avaliar seu estado de degradação, possibilitando novas abordagens de biomonitoramento.

Malaise traps are a common tool for collecting insects used by many researchers. Although there have been variations in the models and materials used for Malaise traps, the potential for sampling bias due to mesh gague has been explored inadequately. This study compared coarse and fine mesh Townes model Malaise traps in two habitats on the Grand Staircase-Escalante National Monument. The two habitats next to the Lick Wash trailhead were defined by dominant vegetation type – sagebrush and grasses or Piñon-Juniper. We collected from three sites per habitat type, over three consecutive days in June in both 2006 and 2007. A pair of Malaise traps consisting of one coarse mesh and one fine mesh trap were used at each site in order to compare differences in the diversity and in the average size of individuals collected by each type of Malaise trap. We measured diversity using both presence-absence data such as richness scores and Jaccard's Index of Similarity, and abundance-based measures of comparison, including Simpson's Index of Diversity and non-metric multidimensional scaling. We identified all individuals according to Order, and because of our interest in flies and their abundance, we further identified the Dipter ta to the Family level. Average insect size was determined by categorizing individuals according to one of 14 distinct size-classes. In sum, 71 samples totaling approximately 62,500 insects were identified and sized. Because we sampled from two adjacent habitats, we also discuss beta diversity across the sample sites. Although mesh-size appears to have a significant effect on the diversity of the catch according to some tests, not all of our analysis agrees. In addition, the gain in the amount of diversity collected by incorporating both mesh-sizes may not be worth the costs of that kind of sampling. Other means of collection may adequately make up that difference. Habitat on the other hand was a clear marker for difference in diversity. Size was not found to be significant overall, but there still may be reasons to examine the effect of mesh-size with respect to the Hymenoptera.

L’industrie microélectronique est un domaine compétitif, confronté de manière permanente à plusieurs défis. Pour évaluer les étapes de fabrication, des tests de qualité sont appliqués. Ces tests étant discontinus, une défaillance des équipements peut causer une dégradation de la qualité du produit. Des alarmes peuvent être déclenchées pour indiquer des problèmes. D’autre part, on dispose d’une grande quantité de données des équipements obtenues à partir de capteurs. Une gestion des alarmes, une interpolation de mesures de qualité et une réduction de données équipements sont nécessaires. Il s’agit dans notre travail à développer des méthodes génériques d’analyse multi-variée permettant d’agréger toutes les informations disponibles sur les équipements pour prédire la qualité de produit en prenant en compte la qualité des différentes étapes de fabrication. En se basant sur le principe de reconnaissance de formes, nous avons proposé une approche pour prédire le nombre de produits restant à produire avant les pertes de performance liée aux spécifications clients en fonction des indices de santé des équipement. Notre approche permet aussi d'isoler les équipements responsables de dégradation. En plus, une méthodologie à base de régression régularisée est développée pour prédire la qualité du produit tout en prenant en compte les relations de corrélations et de dépendance existantes dans le processus. Un modèle pour la gestion des alarmes est construit où des indices de criticité et de similarité sont proposés. Les données alarmes sont ensuite utilisées pour prédire le rejet de produits. Une application sur des données industrielles provenant de STMicroelectronics est fournie
The microelectronics industry is a highly competitive field, constantly confronted with several challenges. To evaluate the manufacturing steps, quality tests are applied during and at the end of production. As these tests are discontinuous, a defect or failure of the equipment can cause a deterioration in the product quality and a loss in the manufacturing Yield. Alarms are setting off to indicate problems, but periodic alarms can be triggered resulting in alarm flows. On the other hand, a large quantity of data of the equipment obtained from sensors is available. Alarm management, interpolation of quality measurements and reduction of correlated equipment data are required. We aim in our work to develop generic methods of multi-variate analysis allowing to aggregate all the available information (equipment health indicators, alarms) to predict the product quality taking into account the quality of the various manufacturing steps. Based on the pattern recognition principle, data of the degradation trajectory are compared with health indices for failing equipment. The objective is to predict the remaining number of products before loss of the performance related to customer specifications, and the isolation of equipment responsible for degradation. In addition, regression- ased methods are used to predict the product quality while taking into account the existing correlation and the dependency relationships in the process. A model for the alarm management is constructed where criticality and similarity indices are proposed. Then, alarm data are used to predict the product scrap. An application to industrial data from STMicroelectronics is provided

Las colecciones de música son cada vez mayores y más variadas, haciendo necesarias nuevas fórmulas para su organización automática. El análisis automático del ritmo tiene como fin la extracción de información rítmica de grabaciones musicales y es una de las principales áreas de investigación en la disciplina de recuperación de la información musical (MIR por sus siglas en inglés). La dimensión rítmica de la música es específica a una cultura y por tanto su análisis requiere métodos que incluyan el contexto cultural. Las complejidades rítmicas de la música clásica de la India, una de las mayores tradiciones musicales del mundo, no han sido tratadas hasta la fecha en MIR, motivo por el cual la elegimos como nuestro principal objeto de estudio. Nuestra intención es abordar cuestiones de análisis rítmico aún no tratadas en MIR con el fin de contribuir a la disciplina con nuevos métodos sensibles al contexto cultural y generalizables a otras tradiciones musicales. El objetivo de la tesis es el desarrollo de técnicas de procesamiento de señales y aprendizaje automático dirigidas por datos para el análisis, descripción y descubrimiento automáticos de estructuras y patrones rítmicos en colecciones de audio de música clásica de la India. Tras identificar retos y posibilidades, así como varias tareas de investigación relevantes para este objetivo, detallamos la elaboración del corpus de estudio y conjuntos de datos, fundamentales para métodos dirigidos por datos. A continuación, nos centramos en las tareas de análisis métrico y descubrimiento de patrones de percusión. El análisis métrico consiste en la alineación de eventos métricos a diferentes niveles con una grabación de audio. En la tesis formulamos las tareas de deducción de metro, seguimiento de metro y seguimiento informado de metro de acuerdo a la tradición estudiada, se evalúan diferentes modelos bayesianos capaces de incorporar explícitamente información de estructuras métricas de niveles superiores y se proponen nuevas extensiones. Los métodos propuestos superan las limitaciones de las propuestas existentes y los resultados indican la efectividad del análisis informado de metro. La percusión en la música clásica de la India utiliza onomatopeyas para la transmisión del repertorio y la técnica. Utilizamos estas sílabas para definir, representar y descubrir patrones en grabaciones de solos de percusión. A tal fin generamos una transcripción automática basada en un modelo oculto de Márkov, seguida de una búsqueda aproximada de subcadenas usando una biblioteca de patrones de percusión derivada de datos. Experimentos preliminares en patrones de percusión de ópera de Pekín, y en grabaciones de solos de tabla y mridangam, demuestran la utilidad de estas sílabas, identificando nuevos retos para el desarrollo de sistemas prácticos de descubrimiento. Las tecnologías resultantes de esta investigación son parte de un conjunto de herramientas desarrollado en el proyecto CompMusic para el mejor entendimiento y organización de la música clásica de la India, con el objetivo de proveer una experiencia mejorada de escucha y descubrimiento de música. Estos datos y herramientas pueden ser también relevantes para estudios musicológicos dirigidos por datos y otras tareas de MIR que puedan beneficiarse de análisis automáticos de ritmo.
Large and growing collections of a wide variety of music are now available on demand to music listeners, necessitating novel ways of automatically structuring these collections using different dimensions of music. Rhythm is one of the basic music dimensions and its automatic analysis, which aims to extract musically meaningful rhythm related information from music, is a core task in Music Information Research (MIR). Musical rhythm, similar to most musical dimensions, is culture-specific and hence its analysis requires culture-aware approaches. Indian art music is one of the major music traditions of the world and has complexities in rhythm that have not been addressed by the current state of the art in MIR, motivating us to choose it as the primary music tradition for study. Our intent is to address unexplored rhythm analysis problems in Indian art music to push the boundaries of the current MIR approaches by making them culture-aware and generalizable to other music traditions. The thesis aims to build data-driven signal processing and machine learning approaches for automatic analysis, description and discovery of rhythmic structures and patterns in audio music collections of Indian art music. After identifying challenges and opportunities, we present several relevant research tasks that open up the field of automatic rhythm analysis of Indian art music. Data-driven approaches require well curated data corpora for research and efforts towards creating such corpora and datasets are documented in detail. We then focus on the topics of meter analysis and percussion pattern discovery in Indian art music. Meter analysis aims to align several hierarchical metrical events with an audio recording. Meter analysis tasks such as meter inference, meter tracking and informed meter tracking are formulated for Indian art music. Different Bayesian models that can explicitly incorporate higher level metrical structure information are evaluated for the tasks and novel extensions are proposed. The proposed methods overcome the limitations of existing approaches and their performance indicate the effectiveness of informed meter analysis. Percussion in Indian art music uses onomatopoeic oral mnemonic syllables for the transmission of repertoire and technique, providing a language for percussion. We use these percussion syllables to define, represent and discover percussion patterns in audio recordings of percussion solos. We approach the problem of percussion pattern discovery using hidden Markov model based automatic transcription followed by an approximate string search using a data derived percussion pattern library. Preliminary experiments on Beijing opera percussion patterns, and on both tabla and mridangam solo recordings in Indian art music demonstrate the utility of percussion syllables, identifying further challenges to building practical discovery systems. The technologies resulting from the research in the thesis are a part of the complete set of tools being developed within the CompMusic project for a better understanding and organization of Indian art music, aimed at providing an enriched experience with listening and discovery of music. The data and tools should also be relevant for data-driven musicological studies and other MIR tasks that can benefit from automatic rhythm analysis.
Les col·leccions de música són cada vegada més grans i variades, fet que fa necessari buscar noves fórmules per a organitzar automàticament aquestes col·leccions. El ritme és una de les dimensions bàsiques de la música, i el seu anàlisi automàtic és una de les principals àrees d'investigació en la disciplina de l'recuperació de la informació musical (MIR, acrònim de la traducció a l'anglès). El ritme, com la majoria de les dimensions musicals, és específic per a cada cultura i per tant, el seu anàlisi requereix de mètodes que incloguin el context cultural. La complexitat rítmica de la música clàssica de l'Índia, una de les tradicions musicals més grans al món, no ha estat encara treballada en el camp d'investigació de MIR - motiu pel qual l'escollim com a principal material d'estudi. La nostra intenció és abordar les problemàtiques que presenta l'anàlisi rítmic de la música clàssica de l'Índia, encara no tractades en MIR, amb la finalitat de contribuir en la disciplina amb nous models sensibles al context cultural i generalitzables a altres tradicions musicals. L'objectiu de la tesi consisteix en desenvolupar tècniques de processament de senyal i d'aprenentatge automàtic per a l'anàlisi, descripció i descobriment automàtic d'estructures i patrons rítmics en col·leccions de música clàssica de l'Índia. Després d'identificar els reptes i les oportunitats, així com les diverses tasques d'investigació rellevants per a aquest objectiu, detallem el procés d'elaboració del corpus de dades, fonamentals per als mètodes basats en dades. A continuació, ens centrem en les tasques d'anàlisis mètric i descobriment de patrons de percussió. L'anàlisi mètric consisteix en alinear els diversos esdeveniments mètrics -a diferents nivells- que es produeixen en una gravació d'àudio. En aquesta tesi formulem les tasques de deducció, seguiment i seguiment informat de la mètrica. D'acord amb la tradició musical estudiada, s'avaluen diferents models bayesians que poden incorporar explícitament estructures mètriques d'alt nivell i es proposen noves extensions per al mètode. Els mètodes proposats superen les limitacions dels mètodes ja existents i el seu rendiment indica l'efectivitat dels mètodes informats d'anàlisis mètric. La percussió en la música clàssica de l'Índia utilitza onomatopeies per a la transmissió del repertori i de la tècnica, fet que construeix un llenguatge per a la percussió. Utilitzem aquestes síl·labes percussives per a definir, representar i descobrir patrons en enregistraments de solos de percussió. Enfoquem el problema del descobriment de patrons percussius amb un model de transcripció automàtica basat en models ocults de Markov, seguida d'una recerca aproximada de strings utilitzant una llibreria de patrons de percussions derivada de dades. Experiments preliminars amb patrons de percussió d'òpera de Pequín, i amb gravacions de solos de tabla i mridangam, demostren la utilitat de les síl·labes percussives. Identificant, així, nous horitzons per al desenvolupament de sistemes pràctics de descobriment. Les tecnologies resultants d'aquesta recerca són part de les eines desenvolupades dins el projecte de CompMusic, que té com a objectiu millorar l'experiència d'escoltar i descobrir música per a la millor comprensió i organització de la música clàssica de l'Índia, entre d'altres. Aquestes dades i eines poden ser rellevants per a estudis musicològics basats en dades i, també, altres tasques MIR poden beneficiar-se de l'anàlisi automàtic del ritme.

En Imagerie hyperspectrale, les volumes de données acquises atteignent souvent le gigaoctet pour une seule et même scène observée. De ce fait, l'analyse de ces données au contenu physique complexe passe obligatoirement par une étape préliminaire de réduction de la dimensionnalité. Cette réduction a un double objectif, le premier consiste à réduire la redondance et le second permet de faciliter les traitements postérieurs (extraction, classification et reconnaissance de formes) et donc l'interprétation des données. La classification automatique est une étape importante du processus d'extraction de connaissances à partir des données. Elle vise à découvrir la structure intrinsèque d'un ensemble d'objets en formant des regroupements qui partagent des caractéristiques similaires. Dans cette thèse, nous nous intéressons à la réduction de dimension dans le cadre de la classification non supervisée des bandes spectrales. Différentes approches existent, comme celles basées sur la projection (linéaire ou non-linéaire) des données de grandes dimensions sur des sous-espaces de représentation bien choisis ou sur les techniques de sélection de bandes spectrales exploitant des critères de complémentarité-redondance d'information qui ne permettent pas de préserver toute la richesse de l'information apportée par ce type de données. 1 - Nous avons accompli une étude comparative, sur la stabilité et la similarité des algorithmes des méthodes non paramétriques et non supervisée de la projection et aussi de la sélection des bandes utilisées dans la réduction de la dimensionnalité à différents niveaux de bruit déterminés. Les tests sont effectués sur des images hyperspectrales, en classant ces derniers en trois catégories selon leur degré de performance de préserver la quantité d'informations. 2 - Nous avons introduit une nouvelle approche de critère basée sur la di-similarité des attributs spectraux et utilisée dans un espace local sur des matrices de données ; L'approche a servi pour définir un taux de préservation d'un évènement rare dans une transformation mathématique donnée. Cependant, nous avons limitée son application au contexte de la thèse liée à la réduction de la taille des données dans une image hyperspectrale. 3 - Les études comparatives ont permis une première proposition d'approche hybride pour la reduction de la taille d'une image hyperspectrale permettant une meilleure stabilité : BandClustering avec Multidimensional Scaling (MDS). Des exemples sont donnés pour démontrer l'originalité et la pertinence de l'hybridation (BandClust / MDS) de l'analyse effectuée. 4 - La tendance de l'hybridation a été généralisée par la suite en présentant un algorithme hybride adaptatif non supervisé basé sur la logique flou (Fuzzy C means), une méthode de projection comme l'analyse en composante principale (ACP) et un indice de validité d'une classification. Les classifications effectuées par Fuzzy C means permettent d'affecter chaque pixel d'une image hyperspectrale à toutes les classes avec des degrés d'appartenance variant entre 0 et 1. Cette propriété rend la méthode FCM intéressante pour la mise en évidence soit des transitions progressives entre les différentes bandes spectrales ou des hétérogénéités spectrales. Grâce à des méthodes conventionnelles appelées indices de validité de classes, nous avons déterminé le nombre optimal de classes de FCM ainsi que le paramètre de flou. Nous montrons que cette hybridation conduit à un taux de réduction pertinent dans l'imagerie hyperspectrale. Par conséquent, Cet algorithme appliqué à différents échantillons de données hyperspectrales, permet une imagerie spectrale beaucoup plus informative, notamment au niveau de l'hétérogénéité spectrale.

Music was the first mass-market industry to be completely restructured by digital technology, and today we can have access to thousands of tracks stored locally on our smartphone and millions of tracks through cloud-based music services. Given the vast quantity of music at our fingertips, we now require novel ways of describing, indexing, searching and interacting with musical content. In this thesis we focus on a technology that opens the door to a wide range of such applications: automatically estimating the pitch sequence of the melody directly from the audio signal of a polyphonic music recording, also referred to as melody extraction. Whilst identifying the pitch of the melody is something human listeners can do quite well, doing this automatically is highly challenging. We present a novel method for melody extraction based on the tracking and characterisation of the pitch contours that form the melodic line of a piece. We show how different contour characteristics can be exploited in combination with auditory streaming cues to identify the melody out of all the pitch content in a music recording using both heuristic and model-based approaches. The performance of our method is assessed in an international evaluation campaign where it is shown to obtain state-of-the-art results. In fact, it achieves the highest mean overall accuracy obtained by any algorithm that has participated in the campaign to date. We demonstrate the applicability of our method both for research and end-user applications by developing systems that exploit the extracted melody pitch sequence for similarity-based music retrieval (version identification and query-by-humming), genre classification, automatic transcription and computational music analysis. The thesis also provides a comprehensive comparative analysis and review of the current state-of-the-art in melody extraction and a first of its kind analysis of melody extraction evaluation methodology.
La industria de la música fue una de las primeras en verse completamente reestructurada por los avances de la tecnología digital, y hoy en día tenemos acceso a miles de canciones almacenadas en nuestros dispositivos móviles y a millones más a través de servicios en la nube. Dada esta inmensa cantidad de música al nuestro alcance, necesitamos nuevas maneras de describir, indexar, buscar e interactuar con el contenido musical. Esta tesis se centra en una tecnología que abre las puertas a nuevas aplicaciones en este área: la extracción automática de la melodía a partir de una grabación musical polifónica. Mientras que identificar la melodía de una pieza es algo que los humanos pueden hacer relativamente bien, hacerlo de forma automática presenta mucha complejidad, ya que requiere combinar conocimiento de procesado de señal, acústica, aprendizaje automático y percepción sonora. Esta tarea se conoce en el ámbito de investigación como “extracción de melodía”, y consiste técnicamente en estimar la secuencia de alturas correspondiente a la melodía predominante de una pieza musical a partir del análisis de la señal de audio. Esta tesis presenta un método innovador para la extracción de la melodía basado en el seguimiento y caracterización de contornos tonales. En la tesis, mostramos cómo se pueden explotar las características de contornos en combinación con reglas basadas en la percepción auditiva, para identificar la melodía a partir de todo el contenido tonal de una grabación, tanto de manera heurística como a través de modelos aprendidos automáticamente. A través de una iniciativa internacional de evaluación comparativa de algoritmos, comprobamos además que el método propuesto obtiene resultados punteros. De hecho, logra la precisión más alta de todos los algoritmos que han participado en la iniciativa hasta la fecha. Además, la tesis demuestra la utilidad de nuestro método en diversas aplicaciones tanto de investigación como para usuarios finales, desarrollando una serie de sistemas que aprovechan la melodía extraída para la búsqueda de música por semejanza (identificación de versiones y búsqueda por tarareo), la clasificación del estilo musical, la transcripción o conversión de audio a partitura, y el análisis musical con métodos computacionales. La tesis también incluye un amplio análisis comparativo del estado de la cuestión en extracción de melodía y el primer análisis crítico existente de la metodología de evaluación de algoritmos de este tipo
La indústria musical va ser una de les primeres a veure's completament reestructurada pels avenços de la tecnologia digital, i avui en dia tenim accés a milers de cançons emmagatzemades als nostres dispositius mòbils i a milions més a través de serveis en xarxa. Al tenir aquesta immensa quantitat de música al nostre abast, necessitem noves maneres de descriure, indexar, buscar i interactuar amb el contingut musical. Aquesta tesi es centra en una tecnologia que obre les portes a noves aplicacions en aquesta àrea: l'extracció automàtica de la melodia a partir d'una gravació musical polifònica. Tot i que identificar la melodia d'una peça és quelcom que els humans podem fer relativament fàcilment, fer-ho de forma automàtica presenta una alta complexitat, ja que requereix combinar coneixement de processament del senyal, acústica, aprenentatge automàtic i percepció sonora. Aquesta tasca es coneix dins de l'àmbit d'investigació com a “extracció de melodia”, i consisteix tècnicament a estimar la seqüència de altures tonals corresponents a la melodia predominant d'una peça musical a partir de l'anàlisi del senyal d'àudio. Aquesta tesi presenta un mètode innovador per a l'extracció de la melodia basat en el seguiment i caracterització de contorns tonals. Per a fer-ho, mostrem com es poden explotar les característiques de contorns combinades amb regles basades en la percepció auditiva per a identificar la melodia a partir de tot el contingut tonal d'una gravació, tant de manera heurística com a través de models apresos automàticament. A més d'això, comprovem a través d'una iniciativa internacional d'avaluació comparativa d'algoritmes que el mètode proposat obté resultats punters. De fet, obté la precisió més alta de tots els algoritmes proposats fins la data d'avui. A demés, la tesi demostra la utilitat del mètode en diverses aplicacions tant d'investigació com per a usuaris finals, desenvolupant una sèrie de sistemes que aprofiten la melodia extreta per a la cerca de música per semblança (identificació de versions i cerca per taral•larà), la classificació de l'estil musical, la transcripció o conversió d'àudio a partitura, i l'anàlisi musical amb mètodes computacionals. La tesi també inclou una àmplia anàlisi comparativa de l'estat de l'art en extracció de melodia i la primera anàlisi crítica existent de la metodologia d'avaluació d'algoritmes d'aquesta mena.

碩士
國立嘉義大學
應用數學系研究所
99
In Ecology, it often uses similarity index to measure the correlation coefficient between two communities of species. The incidence index just considered these species which present or not. However, abundance indices not only considered the presence of species, but also include the information of the relative abundance of species. It can be more explicit to depict the appearance of the entire ecosystems. This study investigated seventeen abundance indices. We try to find out the representative abundance indices through different criteria. First, multidimensional scaling analysis(MDS) and cluster analysis are used to be the criteria for constructing the data structure in the plane space, and observed that whether the migration of species by seasons or not. However, we find that these methods can’t distinguish a valid representative from these indices. Therefore, Hessian matrix and the 3D graph are further used to determine the plane curve of those abundance indices. If the concavity of those indices present convex or concave, suggested to modify their bias. In addition, the concept also applied into incidence indices. Because the properties of linear, invariance and the order of these indices can’t distinguish a representation from these indices, so we give more consideration such as Metric and Euclidean property.

碩士
國立嘉義大學
應用數學系研究所
98
In the ecology, it is the important topic for the compariosn of two assemblages, mainly using the similarity indices to measure their similarity of two assemblages. The similarity indices can be divided into two types, the incidence indices and the abundance indices. The content of this paper is divided into three parts. In the first part, we discuss the range, the linear and invariable properties of incidence indices, and then derive the order of incidence indices. In the second part, we will introduce the way that the abundance indices would be transformed from the incidence indices. And then we discuss the range of the abundance indices and derive the order of these indices. Finally, we will analysis three data set, respectively the data of the bird frequency records of the estuaries of the south region of Hsinchu, the Hong Kong big bird race data, and the data of the seedlings and saplings of in an LEP

碩士
國立成功大學
統計學系碩博士班
95
For comparing two communities, different similarity indices based on species relative abundance data have been proposed in the past. A new class of similarity indices, in which various indices can be defined according to different parameter values, is described in this article. Some of the similarity indices proposed in the past can be included in this class as well, such as Morisita-Horn index, abundance Jaccard index, and abundance Sorensen index. The properties of the proposed indices are described and compared to the conventional indices. Asymptotically consistent estimators of the indices are proposed. The effects of different sample sizes and species diversities on different indices in the proposed index family are illustrated by simulation studies. Also the proposed indices are compared to the conventional ones on the data of successional vegetation in six tropical forests.

碩士
逢甲大學
統計與精算所
96
The estimations of the similar degree between any two communities based on multiple quadrat sampling data are the main purposes of the proposal, such as Morisita-Horn similarity index, Jaccard index, and Sorensen index. The similarity index of any two communities can be measured as an index value between zero and one. We consider the indice that we estimated to be a correlation matrix using the concept that is similar of coefficient correlation. Constructing spatial structure of multiple communities data uses multivariate analysis, and we provide a new deduction of the data and the meaning on space. In order to understand the estimating situation of various indices, we calculate the performances of estimations by simulating several different kinds of species, changed quadrat sampling of counts, and considered if there exists any influence which is made by evolution variation. This analysis which is made of the similarity problem by combining the two concepts of similarity indices and multivariate analysis is more helpful to predict the firsthand information probably. At the final part, we take the materials in soil ciliates species all over the world as the data to discuss primary species distributed situation on Earth.

碩士
逢甲大學
統計與精算所
97
The applicative device of the similarity indices in estimating survival curve in the Jolly-Seber model through open capture-recapture experiment is the main purposes of this proposal. Survival curve is the trend line of the survival rate for a long time, different from the survival rate in Jolly-Seber models. Similarity indices are widely used in the ecological study. Differ from the similarity indices of any two communities, the similarity indices in this proposal are used to derive the estimation of any two strata in different sampling. In open capture-recapture experiment, Jolly-Seber model with its three sub-models are the most popular models in present. Many methods about estimation of survival rate in the Jolly-Seber model were reviewed in the proposal, and these estimations of survival rate are still difficult. And using survival rate to estimate survival curve is also difficult and unsuitable. In order to find a simpler method about estimation of survival curve, we try to analyze the change in similarity indices of any two strata. The estimation of survival curve in the Jolly-Seber model can be derived in a convenient way. We also use the simulation method to check the performance of the estimation, and the Lanyu Scops Owls data is applied in the data analysis. All the work will be stated in the paper.

碩士
國立臺中教育大學
教育測驗統計研究所
102
The purpose of this study is to establish an auto-analyzing system of sentence similarity indices, to explore the trends of “Children corpus” through the 6 indices and furthermore to survey the contribution of the six indices on predicting the text-suitable grade level and the pass rate of the intermediate and high graders on the comprehension test. At last, by combining all the indices which had been established, the study eventually found a readability formula which could be the best predictor. The results were summarized as follows: 1. This study established the indices for Local and Global Minimal Edit Distances of words and part-of-speech(lexical categories), and the sentence syntax similarity between all adjacent sentences and across paragraphs. 2. The contribution of 6 indexes that can predict text-related grade level was up to 22.9% by multiple regression analysis. In terms of stepwise regression analysis, the R2 value made by the 3 indicators (Global Minimal Edit Distances for words、Global Minimal Edit Distances for part-of-speech、Global sentence syntax similarity) was up to 22.7%. 3. Global Minimal Edit Distances of words can predict the pass rate of the 4th,6th graders and all students on the comprehension test in terms of “make straightforward inferences”. The R2 value was up to 23%、17.4%、21.5%. Global Minimal Edit Distances for part-of-speech can predict the pass rate of the 4th, 6th graders and all students on the comprehension test in terms of “interpret and integrate information”. The R2 value was up to 13%、16.8%、15%。 4. The study did a multiple regression analysis with the chosen 48 indices out of 60 in order to predict the text-suitable grade level. The R2 value was up to 80.5%. While through a stepwise regression analysis, the index of Age of acquisition was the best predictor for the text-suitable grade and it produced an R2 value of 74.6%. There were 12 indices got the most notable effect. The R2 value made by the 12 indices was up to 79.8%. 5. If the study did a multiple regression analysis excluding the variable of Age of acquisition from the 48 chosen indices. The other 47 indices could predict the text-suitable grade level with the R2 value as 62.5%. While through a stepwise regression analysis, the R2 value made by the most notable 20 indices was up to 61%. The MTLD was the best index and predictor as well. Mere the index itself could produce an R2 value as 26.4%.

碩士
朝陽科技大學
財務金融系碩士班
97
Many researchers have been endeavoring to analyze and forecast the trend of Taiwan Weighted Stock Indices (TAIEX) from historical data for years. However, the effective duration of the past data applied for the prediction work is still arguable since the future closing prices are not always constituted by the related present closing prices. In this study, the Moving Average approach, both long-term and short-term, is integrated with Genetic Algorithm (GA) and their outcomes are employed as parameters for the grey relational linear moving analysis then others. Four different models including General Fluctuation Average Model, Grey Relational Fluctuation Average Model, General Linear Revision Model, and Grey Relational Linear Revision Model, are implemented to forecasts the trend of TAIEX. The results show that the Grey Relational Linear Model performs higher prediction accuracy. The empirical results are summarized as follows: 1.The best duration parameters acquired from GA for Grey Relational Linear Models are 15-days for short-term and 32-days for long-term respectively. 2.The results of the Globalized Grey Relational Analysis for short-term discovery that comparison sequences prior to or posterior six months are less related to the target sequence. 3.The Grey Relational Fluctuation Average Model and General Linear Revision Model perform better prediction accuracy than that of the General Fluctuation Average Model and General Linear Revision Model in terms of short-term and long-term periods. This result is conformable with our suggestion that historical linear patterns provide effective information for the trend prediction of TAIEX.