Orientador: Michel Daoud Yacoub / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-07T11:29:09Z (GMT). No. of bitstreams: 1 Mendes_JoseRicardo_M.pdf: 915286 bytes, checksum: e4cfda5648e3725df9db2ff028e5c811 (MD5) Previous issue date: 2006 / Resumo: Esta dissertação deduz e analisa coeficientes de correlação de sinais em diversos canais com desvanecimento, bem como especifica os respectivos parâmetros de coerência - distância de coerência, tempo de coerência e largura de banda de coerência. Tais parâmetros são extraídos do coeficiente de correlação da envoltória, estatística fornecida aqui apenas para canais com desvanecimento Rice, devido à complexidade de seu cálculo. Para modelos mais abrangentes de canais com desvanecimento, provê-se o coeficiente de correlação da potência, aproximação simples e geralmente acurada daquele coeficiente. Por fim, através de resultados numéricos, discriminam-se fatores que atuam sobre os parâmetros de coerência: ondas dominantes influem fundamentalmente na distância de coerência e no tempo de coerência; desequilíbrios de potência média dos componentes aleatórios em quadratura, primordialmente na largura de banda de coerência; anisotropia do espalhamento das ondas difusas e diretividade das antenas receptoras, somente na distância de coerência e no tempo de coerência; e a não-estacionariedade do ambiente, em todos os parâmetros de coerência / Abstract: This dissertation derives and examines correlation coefficients of signals in many fading channels, as well as specifies the respective coherence parameters - coherence distance, coherence time, and coherence bandwidth. Such parameters are extracted from the envelope correlation coefficient, statistic calculated here only for Ricean fading channels, owing to the complexity of its deduction. For more comprehensive models of fading channels, it is provided the power correlation coefficient, simple and usually accurate approximation of that coefficient. Finally, by means of numerical results, factors that affect the coherence parameters are enumerated: dominant waves influence fundamentally on the coherence distance and on the coherence time; mean power inequalities between the quadrature random components, mainly on the coherence bandwidth; anisotropy of the diffuse waves spreading, only on the coherence distance and on the coherence time; and nonstationarity of the environment, on all coherence parameters / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

Made available in DSpace on 2014-08-20T14:25:50Z (GMT). No. of bitstreams: 1 dissertacao_maria_arita_garcia.pdf: 8113106 bytes, checksum: fba0a4bf56d450b9bee5e250f53be0b8 (MD5) Previous issue date: 2011-02-18 / This research aims to analyze the monthly variations of rainfall in Rio Grande do Sul (RS) relating to monthly changes in the average thickness of the troposphere (700mb) at the bottom of the Southern Hemisphere, which is centered in the region of Antarctica. Monthly anomalies of atmospheric thickness have been represented by principal oscillation pattern that is called the Arctic Oscillation Index. This study is important because RS has played an important role in national agricultural production, and the variable rainfall has been identified as a key role in the productivity of different regional cultures. During the study, monthly rainfall data from 32 meteorological stations in RS has been used, which has been correlated with the Arctic Oscillation Index. The Arctic Oscillation Index has been obtained directly from Climate Prediction Center and the study period has been from 1979 to 2008 and, the correlations between the monthly rainfalls have been tested considering the time lags of index 0, -1, -2, -3 months. The results of the climatology have showed that rainfall in these last three decades have presented patterns and distributed monthly throughout the year and lower values to the south and higher to the north and northwest. The largest temporal gradient in RS has occurred between the months of October and November, with their maximum and minimum respectively. The Antarctic Oscillation Index has had its greatest variations in the middle of winter and the smallest variations in late summer. The more significant correlations of rainfall in RS and the index have been mostly reversed (by the way the index is built). The highest correlation has been between the rainfall for the month of May and the index also the month of May especially in the northern half of the RS. In the southern half of RS, the changes in the index in March have shown a higher correlation considering rainfall in March and April. The highest correlations in the northwestern part of RS have been happening between the rainfall of November and November Index and the central region of RS has been showing significant correlations in December with the index in November. Among the best results (the highest correlations) it has been highlighted those correlations of the months of March, April and May. From the average rate for the quarter, the extremes of the index have been identified and climatologies of rainfall have been calculated for the two groups (six negative extreme cases and six positive extreme cases). During the negative rainfall period, the index of quarterly rainfall has been exceeding by up to 50% the positive rainfall period, mainly in the south and west of RS. The negative period has been associated to the intensification of high polar consequently it has also been associated to the intensification of the polar jet, thus the most organized cold fronts pass through the RS, while in positive periods the weakening of the high polar and changes in zonal flow of the subtropical jet have been happening, reducing the intensity of cold fronts. / Esta pesquisa tem por objetivo analisar as variações mensais da precipitação no Rio Grande do Sul (RS) e suas relações com as variações mensais da espessura da camada média da troposfera (700mb) na parte inferior do Hemisfério Sul, centrada na região da Antártica. As Anomalias mensais da espessura atmosférica são representadas pelo padrão de oscilação principal, denominado de Índice AAO (Arctic Oscillation Index). A importância deste estudo justifica-se pelo RS ter um papel importante na produção agrícola nacional, sendo a variável precipitação apontada como fator fundamental na produtividade de diversas culturas regionais. Na pesquisa foram utilizados dados de precipitação mensal de 32 estações meteorológicas no RS, as quais foram correlacionadas com o índice AAO. O índice AAO foi obtido diretamente CPC (Climate Prediction Center) e o período de estudo foram de 1979 a 2008, sendo que as correlações entre as precipitações mensais do RS foram testadas com defasagens temporais do índice de 0, -1, -2, -3 meses. Os resultados da climatologia mostram que a precipitação nestas ultimas 3 décadas apresentaram padrões mensais bem distribuídos ao longo do ano, com menores valores ao sul e maiores ao norte e noroeste. O maior gradiente temporal no RS ocorre entre os meses de outubro e novembro, com seus máximos e mínimos respectivamente. O Índice de Oscilação Antártica tem suas maiores variações no meio do inverno e as menores variações no final de verão. As correlações mais significativas da precipitação no RS e o índice foram predominantemente inversas (pela forma como foi construído o índice). As maiores correlações foram entre as precipitações do mês de maio e o Índice também do mês de maio, especialmente na metade norte do RS. Na metade sul do RS, as variações do Índice no mês de março apresentam alta correlação com as precipitações de março e de abril. A parte noroeste do RS as maiores correlações foram entre as precipitações de novembro e o Índice de novembro, Na região central do RS apresentam correlações significativas em dezembro com o Índice de novembro. Entres os melhores resultados (maiores correlações) destacam-se as correlações nos meses de março, abril e maio. A partir do índice médio deste trimestre foram identificados os casos extremos do índice e calculado as climatologias das precipitações para os dois grupos (6 casos extremos negativos e 6 casos extremos positivos). A climatologia mostra que para o período negativo do índice a precipitação acumulada trimestral supera em até 50% a do período positivo, principalmente no sul e oeste do RS. O período negativo está associado à intensificação da alta polar, conseqüentemente intensificação o jato polar e frentes frias mais organizadas passam pelo RS, enquanto que nos períodos positivos ocorre enfraquecimento da alta polar e alteração no fluxo zonal do jato subtropical, reduzindo a intensidade das frentes frias.

Doctor of Philosophy / Department of Statistics / James Higgins / Many studies are performed on units that cannot be replicated due to cost or other restrictions. There often is an abundance of subsampling to estimate the within unit component of variance, but what is needed for statistical tests is an estimate of the between unit component of variance. There is evidence to suggest that the ratio of the between component of variance to the total variance will remain relatively constant over a range of studies of similar types. Moreover, in many cases this intraclass correlation, which is the ratio of the between unit variance to the total variance, will be relatively small, often 0.1 or less. Such situations exist in education, agriculture, and medicine to name a few. The present study discusses how to use such prior information on the intraclass correlation coefficient (ICC) to obtain inferences about differences among treatments in the face of no replication. Several strategies that use the ICC are recommended for different situations and various designs. Their properties are investigated. Work is extended to under-replicated experiments. The work has a Bayesian flavor but avoids the full Bayesian analysis, which has computational complexities and the potential for lack of acceptance among many applied researchers. This study compares the prior information ICC methods with traditional methods. Situations are suggested in which prior information ICC methods are preferable to traditional methods and those in which the traditional methods are preferable.

This article compares the stock exchanges in Vienna, Budapest, Frankfurt and Milan. It settles basic information about their development, the subject of exchange and the classification of market segments. This work also characterizes the trade system of each of the stock exchanges and the liquidity of spot market, with particular emphasis on stock market. A part of this analysis is the comparison of markets based on aspects such as market capitalization, trade volume or the quantity of trade. Last chapter is devoted to the interconnection of stock markets, which is explored by the means of correlation coefficient among different indexes. These indexes show a strong link of the markets and through splitting the timeframe into periods of (2000-2004 - before Hungary's EU entry and after - 2004-2009), a stronger correlation was discovered during the period of index growth (after the Hungary's EU entry). By creating a more autoregressive model VAR, which describes individual processes among stock indexes and the direction of dependency, the hypothesis of strong interconnection of stock markets was proven. VAR model verified one-sided reliance among indexes and the rising level of integration of world markets.

In this thesis, the performance of two over-sampling techniques, SMOTE and ADASYN, is compared. The comparison is done on three imbalanced data sets using three different classification models and evaluation metrics, while varying the way the data is pre-processed. The results show that both SMOTE and ADASYN improve the performance of the classifiers in most cases. It is also found that SVM in conjunction with SMOTE performs better than with ADASYN as the degree of class imbalance increases. Furthermore, both SMOTE and ADASYN increase the relative performance of the Random forest as the degree of class imbalance grows. However, no pre-processing method consistently outperforms the other in its contribution to better performance as the degree of class imbalance varies.

This master’s thesis is engaged in usage of cluster analysis for ECG signal to separate normal QRS complexes from abnormal ones. For this, it is used two algorithms created in professional computing interface MATLAB. The outputs from this master’s thesis are dendrograms, which divide QRS complexes into abnormal and normal clusters, and Pearson correlation coefficients.

Cette thèse concerne l’étude de grandes déviations précises pour deux statistiques de test:le coefficient de corrélation empirique de Pearson et la statistique de Moran.Les deux premiers chapitres sont consacrés à des rappels sur les grandes déviations précises et sur la méthode de Laplace qui seront utilisés par la suite. Par la suite, nous étudions les grandes déviations précises pour des coefficients de Pearson empiriques qui sont définis par:$r_n=\sum_{i=1}^n(X_i-\bar X_n)(Y_i-\bar Y_n)/\sqrt{\sum_{i=1}(X_i-\bar X_n)^2 \sum_{i=1}(Y_i-\bar Y_n)^2}$ ou, quand les espérances sont connues, $\tilde r_n=\sum_{i=1}^n(X_i-\mathbb E(X))(Y_i-\mathbb E(Y))/\sqrt{\sum_{i=1}(X_i-\mathbb E(X))^2 \sum_{i=1}(Y_i-\mathbb E(Y))^2} \, .$. Notre cadre est celui d’échantillons (Xi, Yi) ayant une distribution sphérique ou une distribution gaussienne. Dans chaque cas, le schéma de preuve suit celui de Bercu et al.Par la suite, nous considérons la statistique de Moran $T_n=\frac{1}{n}\sum_{k=1}^n\log\frac{X_i}{\bar X_n}+\gamma \, ,$o\`u $\gamma$, où γ est la constante d’ Euler. Enfin l’appendice est consacré aux preuves de résultats techniques. / This thesis focuses on the study of Sharp large deviations (SLD) for two test statistics:the Pearson’s empirical correlation coefficient and the Moran statistic.The two first chapters aim to recall general results on SLD principles and Laplace’s methodsused in the sequel. Then we study the SLD of empirical Pearson coefficients, name $r_n=\sum_{i=1}^n(X_i-\bar X_n)(Y_i-\bar Y_n)/\sqrt{\sum_{i=1}(X_i-\bar X_n)^2 \sum_{i=1}(Y_i-\bar Y_n)^2}$ and when the meansare known,$\tilde r_n=\sum_{i=1}^n(X_i-\mathbb E(X))(Y_i-\mathbb E(Y))/\sqrt{\sum_{i=1}(X_i-\mathbb E(X))^2 \sum_{i=1}(Y_i-\mathbb E(Y))^2} \, .$ .Our framework takes place in two cases of random sample (Xi, Yi): spherical distributionand Gaussian distribution. In each case, we follow the scheme of Bercu et al. Next, westate SLD for the Moran statistic $T_n=\frac{1}{n}\sum_{k=1}^n\log\frac{X_i}{\bar X_n}+\gamma \, ,$o\`u $\gamma$ , where γ is the Euler constant.Finally the appendix is devoted to some technical results.

In the context of the current energy transition and sustainable development, the promotion of waste separation and waste management is a good means of promoting the energy transition. Waste, when recycled, can contribute to the energy transition in many ways, for example by generating electricity, heating, recycling into raw materials for reuse and producing biofuels. In the process, fossil fuels and overall energy use can be reduced. Sweden is a country with excellent waste management. However, with the current increasing number of university students, many student residences with characteristics that make them isolated from the local community and difficult to manage are emerging. This study was carried out in Visby, Gotland, Sweden. It explored the waste sorting behaviour of students in student accommodation and the extent to which students have knowledge about waste sorting in relation to the energy transition. At the same time, the study attempts to dataset student waste sorting behaviour and knowledge of waste sorting and energy transition, and to find whether there is a relationship between these two variables. The study uses a survey method, and the results are analysed using both qualitative and Pearson’s Correlation test. It was found that the current waste sorting behaviour of students in student accommodation is not very good. This is related to a number of factors, such as whether the students have completed their pre-university studies in Sweden and the students' academic qualifications. According to the survey, the three factors that students think affect their waste sorting behaviour are the benefits of sorting waste, food waste needs to be sorted in special bags and it is difficult to get used to the complex waste sorting system in Sweden. The largest number of respondents felt that it would help them to be more diligent in sorting their waste if waste sorting initiatives could be set up in student residences. The research found a moderate correlation between the waste sorting behaviour of students at Visby and knowledge related to waste sorting and energy transition.

Belastningsskador hos arbetare är ett ökande problem hos olika företag och det har visat sig finnas en tydlig koppling mellan dessa skador och handledens vinkelhastigheten. Det är därför av stort intresse att kunna mäta dessa vinkelhastigheter på ett noggrant och smidigt sätt. Syftet med denna rapport är att utvärdera precisionen av IMU-sensorers förmåga att beräkna vinkelhastigheten av handleden. Detta görs genom att jämföra data från IMU-sensorer med data från ett optiskt spårningssystem (OTS), vilket klassas som en gold standard inom detta område. Ett experiment bestående av åtta övningar utfördes: tre standard rörelser (flexion och rotation i takterna 40, 90 och 140 slag per minut) och fyra simulerade arbeten (målning, pappersvikning, datorarbete och hårföning). Grad av överensstämmelse ges av 1,96 standardavvikelser (SD) för standardrörelserna (10 deltagare) vilka var -31,8 grader/s och 34,2 grader/s, medan för de simulerade arbetena var det -35,1 grader/s och 34,2 grader/s. Det lägsta medelvärdet av medelkvadratavvikelse (RMSD) var 15,7 grader/s och erhölls vid 40 BPM medan den högsta medelvärdet var 93,9 grader/s och erhölls vid målningsövningen. Medelvärdet av korrelationskoefficienten mellan IMU-sensorer och OTS varierade mellan 0,97 och 0,42 och korrelationskoefficienterna av deltagarnas 50:e percentiler av vinkelhastigheten var 0,95 för standardrörelserna och 0,96 för de simulerade arbetena. Medelvärdet av absoluta differensen mellan sensorer och OTS var givet i percentiler (10:e, 50:e och 90:e). Det största spannet för 50:e percentilen gavs vid 140 BPM (18,3 ± 24,6) och det minsta spannet vid 40 BPM (3,5 ± 4,7). Trots att det fanns mindre differenser mellan metodernas mätningar av vinkelhastighet, anser vi att IMU-sensorer har potential att användas för att mäta vinkelhastigheter hos handledens och med vidare utveckling kan den nuvarande differensen minimeras. / Musculoskeletal disorders (MSDs) are increasingly frequent amongst workers and there is a clear connection between work injuries and wrist angular velocities. One of the biggest issues therefore is the currently limited availability of means to measure these angular velocities. The aim of this study is to validate the usability of the IMU sensors to measure angular velocities. This is done by comparing the data from the IMU:s with the data obtained with the optical motion tracking system (OTS), which is considered gold standard within this field of studies. An experiment consisting of eight exercises was conducted: three standard movements (flexion and rotation in the pace 40, 90 and 140 repetitions per minute) and four simulated practical work tasks (painting, folding paper, computer exercise and using a hairdryer). The limits of agreement for the standard movements (10 subjects) were -31,8 degrees/s and 34,2 degrees/s, whereas for the simulated practical work tasks they were -35,1 degrees/s and 28,2 degrees/s. The lowest mean value of the root mean square deviation (RMSD) value was 15,7 degrees/s which represents the 40 BPM task whilst the highest mean value was 93,9 degrees/s which correspond to the painting task. The mean value of the correlation coefficients between the IMU:s and the OTS ranged between 0,97 and 0,42 and the correlation coefficient between the subjects 50:th percentiles of the angular velocity, was 0,95 for the standard movements whilst for the practical work tasks it was 0,96. The mean value of the absolute difference between the sensors and the OTS was given in percentiles (10th, 50th and 90th). The largest range within the 50th percentile occurred during the 140 BPM task (18,3 ± 24,6) and the smallest range during the 40 BPM task (3,5 ± 4,7). Although the measured angular velocities vary to a certain extent between the two methods, we conclude that the IMU sensors present the potential to work as measuring units for wrist angular velocities and with further development the current differences can be minimized. / Forte dnr: 2017-01209 "Enkel och tideffektiv metod att mät, analysera och presentera biomekaniskbelastning för hand-handled"

This thesis investigates imbalanced Swedish text financial datasets, specifically receipt classification using machine learning models. The study explores the effectiveness of under-sampling and over-sampling methods for Naive Bayes algorithms, collaborating with Fortnox for a controlled experiment. Evaluation metrics compare balancing methods regarding the accuracy, Matthews's correlation coefficient (MCC) , F1 score, precision, and recall. Findings contribute to Swedish text classification, providing insights into balancing methods. The thesis report examines balancing methods and parameter tuning on machine learning models for imbalanced datasets. Multinomial Naive Bayes (MultiNB) algorithms in Natural language processing (NLP) are studied, with potential application in image classification for assessing industrial thin component deformation. Experiments show balancing methods significantly affect MCC and recall, with a recall-MCC-accuracy tradeoff. Smaller alpha values generally improve accuracy.  Synthetic Minority Oversampling Technique  (SMOTE) and Tomek's algorithm for removing links developed in 1976 by Ivan Tomek. First Tomek, then SMOTE (TomekSMOTE)  yield promising accuracy improvements. Due to time constraints, Over-sampling using SMOTE and cleaning using Tomek links. First SMOTE, then Tomek (SMOTETomek) training is incomplete. This thesis report finds the best MCC is achieved when $\alpha$ is 0.01 on imbalanced datasets.