Search

Mothers’ preference for introducing and elaborating on mental states in conversations with their young children has been shown to be a significant predictor of children’s Theory of Mind development among Western samples. However, it is unclear whether this relationship still holds true in the Chinese cultural context. Exploratory in nature, the present study aims to study the relationship between mothers’ mental state talk and the development of Theory of Mind (ToM) in Chinese children (4-5 years old) following the framework of Peterson and Slaughter’s (2003) research. All participants (N = 94) were from middle-income families and were Chinese mothers of a 4-5-year-old typically developing children. Fifty-two pre-schoolers (M = 4.47 years, SD = .49) were boys. Adopting a correlational study design, the current study first looks at if there were individual differences in mothers’ conversational styles with their children. By running correlational analyses examining the relationships between different conversational styles, this assumption was confirmed. Given individual differences did exist, Pearson’s correlation was performed to decide if mothers’ preferences of language use significantly correlated with ToM development in Chinese children. The result from the correlation analysis indicated that no significant correlation emerged between mothers' self-reported preferences for sophisticated explanations of mental states and parental report of children's ToM. This null result was again confirmed by a hierarchical regression analysis showing that mothers’ conversational preferences were not predictors of Chinese children’s ToM development when relevant control variables were included in the analyses. Additionally, two correlational analyses were conducted to address the research question that whether children’s age and expressive vocabulary correlate with parental reports of ToM understandings. No correlation was found between age and ToM ability. However, there was a significant correlation between children’s expressive language and ToM, giving further support to the social constructivist account. An independent samples t-test to test the gender effect on parental-report ToM yielded no significant results. Evidence of Internal consistency, content validity, and construct validity for the two translated questionnaires was given. However, accurate estimation of the psychometric properties of the two translated measures requires more research. Finally, Implications and future directions for research on the relationship between maternal mental state talk and ToM development in diverse and larger samples from a longitudinal lens were discussed.

Principal component analysis (PCA) and kernel PCA allow the decorrelation of data with respect to a basis that is found via variance maximization. However, these techniques are based on pointwise correlations. Especially in the context of time series analysis this is not optimal. We present a novel generalization of PCA that allows to imprint any desired correlation pattern. Thus the proposed method can be used to incorporate previously known statistical dependencies between input variables into the model which is increasing the overall performance. This is achieved by generalizing the projection onto the direction of maximum variance—as known from PCA—to a projection onto a multi-dimensional subspace. We focus on the use of cyclic correlation patterns, which is especially of interest in the domain of time series analysis. Beneath introducing the presented variation of PCA, we discuss the role of this method with respect to other well-known time series analysis techniques., publishedVersion

Background Older adults living with frailty who require treatment in hospitals are increasingly seen in the Emergency Departments (EDs). One quick and simple frailty assessment tool—the Clinical Frailty Scale (CFS)—has been embedded in many EDs in the United Kingdom (UK). However, it carries time/training and cost burden and has significant missing data. The Hospital Frailty Risk Score (HFRS) can be automated and has the potential to reduce costs and increase data availability, but has not been tested for predictive accuracy in the ED. The aim of this study is to assess the correlation between and the ability of the CFS at the ED and HFRS to predict hospital-related outcomes. Methods This is a retrospective cohort study using data from Leicester Royal Infirmary hospital during the period from 01/10/2017 to 30/09/2019. We included individuals aged + 75 years as the HFRS has been only validated for this population. We assessed the correlation between the CFS and HFRS using Pearson’s correlation coefficient for the continuous scores and weighted kappa scores for the categorised scores. We developed logistic regression models (unadjusted and adjusted) to estimate Odds Ratios (ORs) and Confidence Intervals (CIs), so we can assess the ability of the CFS and HFRS to predict 30-day mortality, Length of Stay (LOS) > 10 days, and 30-day readmission. Results Twelve thousand two hundred thirty seven individuals met the inclusion criteria. The mean age was 84.6 years (SD 5.9) and 7,074 (57.8%) were females. Between the CFS and HFRS, the Pearson correlation coefficient was 0.36 and weighted kappa score was 0.15. When comparing the highest frailty categories to the lowest frailty category within each frailty score, the ORs for 30-day mortality, LOS > 10 days, and 30-day readmission using the CFS were 2.26, 1.36, and 1.64 and for the HFRS 2.16, 7.68, and 1.19. Conclusion The CFS collected at the ED and the HFRS had low/slight agreement. Both frailty scores were shown to be predictors of adverse outcomes. More research is needed to assess the use of historic HFRS in the ED.

El aumento de la inteligencia y la capacidad de cálculo de los equipos de usuario (UE) deja la puerta abierta a un nuevo paradigma, en el que los dispositivos personales toman un papel activo en su integración como parte de la RAN. El uso de estas UE como relés es la tendencia dominante, ayudando a otros usuarios al enlace en la estación base cuando experimentan condiciones de mala cobertura. En este contexto, el trabajo desarrollado en este proyecto pretende contribuir y discutir si el UE puede ser consciente de una situación adversa cercana, es decir, pérdida de conexión o handover, estudiando y aplicando técnicas de análisis de series temporales en secuencias reales medidas en células 5G. Los resultados indican que las grabaciones que siguen el mismo camino por lo general tienen suficientes semejanzas para relacionarlos, lo que significa que la historia de un UE anterior se puede utilizar para avanzar una situación que puede empeorar. L'augment de la intel·ligència i la capacitat de càlcul dels equips d'usuari (UE) deixa la porta obert a un nou paradigma, on els dispositius personals prenen un paper actiu en la seva integració com a part de la RAN. L'ús d'aquestes UE com a relés és la tendència dominant, ajudant a altres usuaris a l'enllaç a l'estació base quan experimenten condicions de mala cobertura. En aquest context, el treball desenvolupat en aquest projecte pretén contribuir i discutir si el UE pot ser conscient d'una situació adversa propera, és a dir, pèrdua de connexió o handover, estudiant i aplicant tècniques d'anàlisi de sèries temporals en seqüències reals mesurades en cèl·lules 5G. Els resultats indiquen que els enregistraments que segueixen el mateix camí en general tenen prou semblances per relacionar-los, el que significa que la història de un UE anterior es pot utilitzar per avançar una situació que pot empitjorar. The increased intelligence and computing capacity of user equipment (UE) leaves the door open to a new paradigm, where personal devices take an active role in being integrated as a part of the RAN. The use of these UEs as relays is the dominant trend, assisting other users in the link to the base station when they experience poor coverage conditions. In this context, the work developed in this project is intended to contribute to and discuss if the UE can be aware of an upcoming adverse situation i.e., loss of connection or handover, studying and applying time series analysis techniques in real sequences measured in 5G cells. The findings indicate that recordings following the same path generally have enough similarities to relate them, meaning that previous UE’s history can be used to notice in advance a worsening situation.

This article explores U-Statistics as a tool for testing conditional correlation between two multivariate sources with respect to a potential confounder. The proposed approach is effectively an instance of weighted U-Statistics and does not impose any statistical model on the processed data, in contrast to other well-known techniques that assume Gaussianity. By avoiding determinants and inverses, the method presented displays promising robustness in small-sample regimes. Its performance is evaluated numerically through its MSE and ROC curves.

Treballs Finals del Grau de d'Administració i Direcció d'Empreses, Facultat d'Economia i Empresa, Universitat de Barcelona, Curs: 2021-2022, Tutor: Alberto Heras Oliver, La relació entre l’administració d’una empresa i la Fórmula 1 té una importància creixent, ara que aquesta gosa de cada cop més èxit i popularitat, sobretot en les audiències joves. Al cap i a la fi, un equip esportiu no deixa de ser una empresa. Aquest treball se centra a determinar quin és el factor que té més pes per una empresa a l’hora d’entrar a la competició i invertir-hi grans quantitats de diners. Quina és la motivació real per la qual les marques hi participen. L’estudi consisteix en recol·lectar dades i tractar-les per tal de poder-les estudiar amb un sentit lògic, i calcular a partir d’aquesta informació uns indicadors estadístics que permetin projectar uns resultats. Paral·lelament, s’estudia i s’aprofundeix en el funcionament de la Fórmula 1 i el seu entorn per poder-ne extreure unes conclusions sòlides i fiables que responguin als objectius plantejats.

Treballs Finals del Doble Grau d'Administració i Direcció d'Empreses i de Matemàtiques, Facultat d'Economia i Empresa i Facultat de Matemàtiques i Informàtica, Universitat de Barcelona, Curs: 2021-2022, Tutor: Josep Vives i Santa Eulàlia i F. Javier Martínez de Albéniz, [en] Time series analysis allows complex processes to be expressed in simple terms to understand how these processes were generated and to predict future values. SARIMA models assume that the observations of a process depend on the previous observations and the variation between them to give an expression of the underlying data generating process. To find the SARIMA model that better fits our data we introduce the Box and Jenkins method, based on three iterative steps: model identification, parameter estimation and fitness check. Once we have identified the most appropriate fitting model, we use it to forecast future values. We have followed this methodology to find the model that best fits the Spanish unemployment series from 2002 to the first quarter of 2022 and to forecast the next 8 observations.

Treballs Finals del Màster de Ciències Actuarials i Financeres, Facultat d'Economia i Empresa, Universitat de Barcelona, Curs: 2021-2022, Tutor: Miguel Ángel Santolino Prieto, En el presente trabajo se analizará la serie temporal de exceso de mortalidad provocado por el Covid-19 en España, diferenciando entre sexos y franjas de edades. Se escogerán también cuatro series financieras, las cuales se analizarán y trabajarán para que cumplan las condiciones necesarias para poder aplicar modelos estadísticos. Entre las series de exceso de mortalidad y las series financieras se analizará su posible correlación cruzada y si tienen relación de causalidad detectable mediante el Test de Granger. La relación de causalidad nos indicará si la serie de exceso de mortalidad es buena predictora para la serie financiera.

Treballs Finals del Màster de Ciències Actuarials i Financeres, Facultat d'Economia i Empresa, Universitat de Barcelona, Curs: 2021-2022, Tutor: Samer Ajour El Zein, El presente trabajo consiste en un análisis del impacto de las noticias de prensa sobre la volatilidad de Bitcoin. Se realiza un análisis gráfico tomando una base de datos de noticias importada de un importante periódico económico nacional, y se observa el impacto de ciertas palabras clave sobre la serie de precios de Bitcoin. El análisis se complementa con un análisis estadístico basado en los modelos GARCH, y concretamente utilizando el modelo multivariante DCC. Este modelo muestra que las variables de estudio se correlacionan con la variable principal, concluyendo que sí es posible estimar el impacto que pueden tener las noticias sobre esta cryptomoneda.

Treballs Finals del Grau de d'Administració i Direcció d'Empreses, Facultat d'Economia i Empresa, Universitat de Barcelona, Curs: 2021-2022, Tutor: Miguel Ángel Santolino Prieto, Aquest treball de fi de grau se centra en l’anàlisi de la mortalitat durant els anys de pandèmia Covid-19 i la seva comparativa amb l’IBEX-35 així com el seu efecte en diversos sectors de l’economia espanyola. Per dur a terme la primera part del treball s’han calculat els excessos de mortalitat setmanals a Espanya dels anys 2020 i 2021 respecte la mitjana dels 19 anys anteriors. A continuació s’ha procedit amb un anàlisi de correlacions entre la mortalitat i el comportament de l’IBEX-35. Finalment s’ha dut a terme un anàlisi de correlacions de l’evolució de la mortalitat amb l’evolució a borsa de les empreses Meliá Hotels International, S.A i Pharma Mar, S.A representants dels sectors turístic i farmacèutic respectivament. D’aquesta segona part es conclou que existeix una correlació entre l’excés de mortalitat de la pandèmia i el comportament de la borsa espanyola.

Orientador: Lireny Ap. G. Gonçalves Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia de Alimentos Resumo: DiaMed Food Analysis Test System (F.A.T.S.) é composto por testes rápidos, alternativos aos métodos convencionais de determinação de ácidos graxos livres (%AGL), índice de peróxidos (IP), concentração de alquenais (teste de p-anisidina) e malonaldeído (teste de TBA). Esses testes podem ser empregados na avaliação de óleos, gorduras e produtos que os contenham. Além disso, os equipamentos são compactos, não necessitando de laboratórios altamente equipados e as quantidades relativamente pequenas de amostra requeridas geram menos resíduos. Eles oferecem, ainda, melhores condições de trabalho aos analistas envolvidos do que os métodos convencionais. Substituem, por exemplo, o reagente de p-anisidina, considerado carcinogênico. Até o momento, apenas os kits de %AGL e IP foram certificados pela AOAC e existem poucos estudos utilizando-os. Em vista disso, o presente trabalho teve como objetivo correlacionar os kits com os métodos oficiais da AOCS, aplicados a óleos vegetais brutos e degomados, óleos refinados, azeites de oliva (avaliados com 4 kits de diferentes faixas de detecção de %AGL e IP), óleos de fritura coletados no restaurante universitário (TBA e p-anisidina) e rações fornecidas por empresa nacional (%AGL, IP, TBA e p-anisidina). Em todos os casos, as análises foram conduzidas em triplicata, utilizando os métodos convencionais. No dia seguinte, analisaram-se as mesmas amostras com os kits correspondentes, de acordo com as instruções do fornecedor. Adotaram-se os métodos Ca 5a-40, Cd 8b-90, Cd 18b-90 e Cd 19b-90, respectivamente, para %AGL, IP, teste de p-anisidina e TBA. As determinações para as amostras de ração foram feitas no lipídio extraído a frio com mistura de éter de petróleo e éter etílico. A correlação entre metodologias foi determinada pelo Método dos Mínimos Quadrados de Regressão Linear (usando o programa Minitab for Windows versão 12.1) e o teste de ANOVA foi utilizado para comparar médias de diferentes metodologias (software SAS for Windows V. 8). Para as amostras de ração e óleo de fritura, a variação entre as amostras testadas usando um mesmo método foi também determinada. Houve alta correlação entre os resultados do método oficial e pelo kit em relação à %AGL e IP de óleos refinados e azeites de oliva e à %AGL das rações e aos valores de p-anisidina de óleos de fritura (r = 0,74-0,99). Os resultados de óleos brutos e degomados foram significativamente diferentes, devido à interferência dos pigmentos presentes. As amostras de ração apresentaram estágio oxidativo pouco avançado, de forma que não puderam ser avaliadas precisamente com os kits. Para os óleos de fritura, o kit de TBA foi mais sensível que o método oficial. Além disso, constataram-se deficiências nos métodos oficiais de determinação de IP, malonaldeído e alquenais, uma vez que o método de peróxido não foi sensível para IP menor que 2 meq/kg, ao passo que o aquecimento empregado no teste de TBA gerou valores mais elevados que os obtidos com o kit equivalente e a presença de água nas amostras e reagentes interferiram nos resultados de p-anisidina. Sendo assim, os kits DiaMed F.A.T.S. forneceram resultados precisos e se tornaram uma opção viável para as superar as limitações econômicas dos estudos de vida útil enfrentadas pelos métodos convencionais, especialmente para amostras com baixo teor lipídico Abstract: The DiaMed Food Analysis Test System (F.A.T.S.) consists of rapid tests which serve as alternatives for conventional methods in the determination of free fatty acids (%FFA), peroxide value (PV), and the concentration of alkenal (p-anisidine test) and malonaldehyde (TBA test). These tests can be employed in the evaluation of oils, fats and products containing them. Moreover, the compact equipment makes a large, well-equipped laboratory unnecessary and the relatively small sample size generates minimum residues. They also furnish better working conditions for the analysts involved since the p-anisidine used as a reagent is carcinogenic. At present, only the %FFA and PV kits have been certified by the AOCS, but few studies have used them. The present study was thus designed to correlate the results obtained from the use of these kits with those obtained by the official AOCS methods for crude, degummed and refined vegetable oils and olive oils (tested with 4 kits for different levels of %FFA and PV), frying oils collected from the university restaurant (tested for TBA and p-anisidine values) and pet foods supplied by a national industry (tested for %FFA, PV and TBA and p-anisidine values). In all cases the analyses were performed in triplicate using conventional methods. On the following day, the same samples were analyzed using the corresponding kits, according to the instruction of the manufacturer. The official AOCS methods adopted for the determination of %FFA, PV and p-anisidine and TBA values were Ca 5a-40, Cd 8b-90, Cd 18b-90 and Cd 19b-90, respectively. For the pet food samples, lipids were extracted with a mixture of petroleum ether and ethylic ether in cold conditions. The correlations between methodologies were determined using the Least Squares Method of Linear Regression (using Minitab for Windows, version 12.1), and an ANOVA was used to compare the averages of the different methods (SAS for Windows V. 8 software). For both pet food and frying oil samples, the variation between samples tested using the same method was also determined. High correlations were found between the results of official and kit methods in relation to %FFA and PV for refined and olive oils, and in relation to %FFA in pet foods and p-anisidine values in frying oils (r = 0.74-0.99). Results for crude and degummed oils were significantly different, however, due to the effect of pigments. The pet foods exhibited a relatively low oxidative state and thus could not be analyzed accurately with the kits. For frying oils, the TBA kit proved more sensitive than the official method. Moreover, the official methods presented problems for the determination of PV and malonaldehyde and alkenal concentrations, since they were not very sensitive for PV levels of less than 2 meq/kg, whereas the use of heating in the official TBA test led to higher values than those obtained when using the equivalent kit and the presence of water in the samples and reagents interfered in the results of the p-anisidine test. The DiaMed F.A.T.S. kits thus proved to provide accurate results. Moreover, they provide a feasible option for overcoming the economic difficulties of shelf-life studies faced by conventional methods, especially for samples with a low lipid content Mestrado Tecnologia de Alimentos Mestre em Tecnologia de Alimentos

It is well known that Dengue is a virus mainly spread by mosquito, the Aedes Aegypty, that is common in the area of Brazil. This country had been very affected by dengue disease, as well as ZiKa and chicunguña. This mosquito was originated in Africa, but now is found in tropical, subtropical and temperate regions throughout the world. Climatological conditions are very related with the number of mosquitoes and the viruses spread. The availability of Earth Observation mission that periodically and globally provide of climatological variables such as temperature and precipitations, may help to develop dengue prevention algorithms. The passive remote sensing group at Signal theory and propagation group lately had started to work in the development of such algorithms. The main objective of this thesis has been the improvement of artificial intelligence algorithms for dengue risk detection in Brazil. For achieving such objective, the original database has been upgraded adding new years of information, with all the information: monthly dengues cases in in Brazil, as well as all the climatologic and socioeconomic variables associated and necessary for the dengue risk prediction. In addition, the predictive algorithms based in Machine Learning and Deep Learning already developed in a previous project had been adapted with the new information in the database, as well as analyzed and improved. This document details the information of the database, its analysis and explains the different statistical models that have been used to make prediction of the epidemiological risk of dengue in a month ahead. Moreover, an analysis of the performances of each model and a comparative study among them is presented. Es bien conocido que el Dengue es un virus principalmente transmitido por el mosquito, Aedes Aegypty, que es común en la zona de Brasil. Este país ha sido muy afectado por esta enfermedad, juntamente con el Zika i la Chikunguña. El mosquito era originario del África, pero ahora se encuentra en áreas tropicales, subtropicales y temperadas de todo el mundo. Las condiciones meteorológicas están muy relacionadas con el número de mosquitos, y por tanto con la propagación de Dengue. El fácil acceso a la misión "Earth Observation" que periódicamente proporciona a nivel mundial variables meteorológicas como la temperatura i las precipitaciones, ha ayudado al desarrollo de algoritmos de prevención del Dengue. El grupo de obtención de datos pasivos al grupo de teoría de señal i propagación del virus ha empezado a trabajar en la implementación de este. El principal objetivo de este proyecto ha estado la mejora de los algoritmos de inteligencia artificial encargados de la predicción de dengue en Brasil. Para cumplir este objetivo se ha mejorado la base de datos original añadiendo más años de información, con toda esta información: el número de casos mensuales de dengue en Brasil, también datos climatológicos i socioeconómicos relacionados i necesarios para la predicción de dengue. A más a más, los algoritmos predictivos basados en Machine Learning i Deep Learning que ya han sido implementados en proyectos anteriores han sido adaptados con la nueva información de la base de datos, también analizados y mejorados. Este documento detalla la información de la base de datos, analiza esta y explica los diferentes modelos estadísticos utilizados para hacer una predicción futura del riesgo epidemiológico del dengue. Adicionalmente, para cada modelo se ha hecho un análisis de su rendimiento y se ha hecho un estudio para comparar los modelos. És ben conegut que el Dengue és un virus principalment transmès per el mosquit, l’Aedes Aegypty, que és comú a la zona de Brasil. Aquest país ha estat molt afectat per aquesta malaltia, juntament amb el Zika i la chicuguña. El mosquit era originari de l’Africa, però ara es troba en àrees tropicals, subtropicals i temperades de tot el món. Les condicions meteorològiques estan molt relacionades amb el nombre de mosquits, i per tant amb la propagació del dengue. El fàcil accés a la missió “Earth Observation” que periòdicament proporciona a nivell mundial variables meteorològiques com la temperatura i les precipitacions, ha ajudat a desenvolupar algorismes de prevenció del Dengue. El grup de obtenció de dades passives al grup de teoria de senyal i propagació del virus ha començat a treballar en la implementació d’aquests. El principal objectiu d’aquest projecte ha estat la millora dels algorismes d’intel·ligència artificial encarregats de la predicció de dengue al Brasil. Per tal d’assolir aquest objectiu s’ha millorat la base de dades original afegint més anys d’informació, amb tota la informació: els casos mensuals de dengue a Brasil, també dades climatològiques i socioeconòmiques relacionades i necessàries per la predicció de dengue. A més a més, els algorismes predictius basats en Machine Learning i Deep Learning que ja han estat implementats en projectes anteriors han estat adaptats amb la nova informació de la base de dades, també com analitzats i millorats. Aquest document detalla la informació de la base de dades, l’anàlisi d’aquesta i explica els diferents models estadístics utilitzats per fer una predicció a un mes vista del risc epidemiològic del dengue. Addicionalment, per cada model s’ha fet un anàlisi del seu rendiment i s’ha fet un estudi per comparar els diferents models.

Informe intern UPC - UAB - ICTA Mentre el nombre de correlacions positives és molt superior que les negatives per la pluja hivernal que per l'anual a la majoria d’estacions aerobiològiques estudiades (Barcelona, Tarragona, Lleida i Manresa), a Bellaterra el nombre de correlacions negatives supera al de positives, és a dir, la raó positives/negatives és inferior a 1, igual que passava amb la pluja anual. A Girona aquesta raó també és inferior a 1, però el nombre de correlacions positives i negatives està molt més equilibrat. En aquest informe s’exposa un anàlisi més detallat del què passa a totes dues estacions, amb la intenció d’explicar aquest comportament inesperat, especialment a Bellaterra.

Orientador: Verónica Andrea González-López Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática Estatística e Computação Científica Resumo: O objetivo dessa pesquisa foi o de analisar variáveis que compõem o indicador da qualidade de educação do estado de São Paulo (IDESP) na cidade de Guarulhos. A principal justificativa para esse trabalho é de visualizar ano a ano os resultados das escolas sem real evolução, repetindo os mesmos resultados alcançados em outras edições. O questionamento central é de que pode existir certa manipulação em uma dessas variáveis para implicar em melhores resultados. Após um breve resumo sobre a implementação do sistema de avaliação em larga escala no Brasil, apresentamos a composição do IDESP e a distribuição de seus resultados na cidade de Guarulhos nos anos de 2.013, 2.014 e 2.015, bem como os principais valores descritivos de cada ano junto a uma análise de seus resultados que serão em seguida validadas através do teste de correlação de Spearman. Para uma melhor compreensão sobre a dependência entre os componentes que compõem o IDESP adotamos a modelagem entre um determinado componente do IDESP, que é um indicativo de mal desempenho escolar, e a taxa de reprovação, utilizando a cópula Secções Cúbicas Assimétricas (ACS) para descrever a dependência, onde podemos comprovar diferentes comportamentos na relação de dependência para o ensino fundamental e médio entre os anos. Entre as conclusões damos destaque à falta de coerência percebida na dependência entre dois componentes do IDESP que por uma visão simplista deveriam ser concordantes, mas apresentam resultados que destoam do esperado Abstract: The goal of this research is to analyze variables that compose the indicator of the quality of education of the state of São Paulo (IDESP) in the city of Guarulhos. The main justification for this work is to visualize year after year the results of schools without achieving a real evolution, repeating the same results achieved in other editions. After a brief summary on the implementation of the large-scale evaluation system in Brazil, we present the composition of IDESP and the distribution of its results in the city of Guarulhos in the years of 2013, 2014 and 2015, as well as the main descriptive values of each year together with an analysis of its results, which will then be validated through the Spearman correlation test. For a better understanding of the dependence between the components that compose the IDESP adopted the modeling between a certain component of the IDESP, which is an indicative of poor school performance and the rate of reprobation using the Asymmetric Cubic Sections (ACS) family of Copulas describe the dependence, we can verify different types of dependences. Among the most outstanding conclusions, we see that the results show that the dependence found is not always reasonable, so we suspect that there is some kind of interference in the construction of these variables Mestrado Matemática em Rede Nacional Mestre em Matemática CAPES 5512227

Orientador: Renato da Rocha Lopes Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação Resumo: Nesta dissertação, nós discutimos o cálculo de funções de coerência de alta resolução para a estimação dos parâmetros de empilhamento em processamento de sinais sísmicos. Nosso foco é o algoritmo de estimação por Classificação de Sinais Múltiplos (MUSIC, do inglês MUltiple SIgnal Classification). Este pode superar a tradicional função de coerência semblance em casos em que há reflexões próximas ou interferentes. Nossa principal contribuição é a proposta de diversas simplificações para sua implementação. Primeiro, mostramos como obter os valores da função MUSIC a partir do subespaço de sinais do dado sísmico, que possui dimensão menor do que o subespaço de ruído, usualmente empregado. Depois disso, mostramos como obter o subespaço de sinais a partir do método da potência. Chamamos esta técnica de MUSIC com Método da Potência (PM-MUSIC). Também propusemos uma nova maneira de obtenção do espectro de MUSIC, baseada na decomposição em autovalores e autovetores da matriz de correlação temporal do dado sísmico. Este método contrasta com os algoritmos presentes na literatura, que se baseiam na correlação espacial. A partir do uso do Método da Potência, obtivemos reduções de complexidade tanto para a variante espacial quanto para a temporal do algoritmo MUSIC. Finalmente, também propusemos uma nova função de normalização para o cálculo de MUSIC, a qual chamamos de ponderação por semblance. Esta função leva em conta o espectro de velocidades obtido com a função de coerência semblance e lida com a alta variação dinâmica produzida pelo espectro de velocidades calculado com MUSIC. Nós comparamos a implementação de PM-MUSIC, a partir das correlações temporal e espacial. Exemplos numéricos com dados sísmicos sintéticos e de levantamentos reais demonstraram que o algoritmo PM-MUSIC supera o semblance e que sua variante temporal possui alta resolução, assim como sua variante espacial. Além disso, PM-MUSIC obtido a partir da correlação temporal mostrou-se extremamente robusto ao lidar com sinais correlacionados Abstract: In this dissertation we discuss high-resolution coherence functions for the estimation of the stacking parameters in seismic signal processing. We focus on the MUltiple SIgnal Classification (MUSIC) algorithm, which uses the eigendecomposition of the seismic data to measure the coherence. MUSIC can outperform the traditional semblance in cases of close or interfering reflections. Our main contribution is to propose several simplifications to the implementation of MUSIC. First, we show how to compute MUSIC coherence measure in terms of the signal subspace of seismic data, which has lower dimension than the one currently used, the noise subspace. After that, we show how to obtain the signal subspace, iteratively, with the power method. We called this technique of Power Method MUSIC (PM-MUSIC). We also propose a new way to obtain the MUSIC spectrum, based on the eigendecomposition of the temporal correlation matrix of the seismic data. This is in contrast to the algorithms in the literature, which are based on the spatial correlation. Complexity reductions are obtained and discussed with the use of Power Method for both spatial and temporal variant of MUSIC. Finally, we propose a new normalization function for MUSIC, which we called semblance weighting. This function takes into account semblance coefficient and deals with high dynamic range in MUSIC velocity spectrum. We compared spatial and temporal correlation matrices, implemented with PM-MUSIC. Numerical examples with synthetic and real seismic data indicated that PM-MUSIC outperforms semblance and that the temporal variant of PM-MUSIC can present the same high-resolution as its spatial counterpart. Moreover, temporal PM-MUSIC is particularly useful when dealing with correlated signals Mestrado Telecomunicações e Telemática Mestre em Engenharia Elétrica

Treballs Finals de Grau en Estadística UB-UPC, Facultat d'Economia i Empresa (UB) i Facultat de Matemàtiques i Estadística (UPC), Curs: 2020-2021, Tutor: Ernest Pons Fanals, En aquest projecte es durà a terme a partir d’unes dades obtingudes a través de l’Oficina d’Accés a la Universitat, en concret de l’any 2016. Aquestes dades corresponen als estudiants que han sol·licitat plaça en alguna de les universitats públiques de Catalunya. L’objectiu del treball es centra en analitzar els estudiants que habiten a Barcelona. S’estudiarà la possible influència d’un seguit de variables, tant socio-demogràfiques com econòmiques, en la decisió de la branca d’estudi de cada individu. Una de les variables que causa més interès és el barri o districte en el que habiten. A partir d’un seguit d’anàlisi descriptius i de tests estadístics, s’arribarà a una conclusió sobre quins són els factors contemplats que més relació poden tenir en aquest procés.

Treballs Finals del Màster de Ciències Actuarials i Financeres, Facultat d'Economia i Empresa, Universitat de Barcelona, Curs: 2020-2021, Tutor: Marta Gómez Puig, El objetivo de este trabajo es analizar si la Unión Económica y Monetaria Europea (UEM) cumple los requisitos principales para utilizar una moneda única y, formar un Área Monetaria Óptima. Según la teoría económica, uno de los requisitos más importantes es la sincronía de los ciclos económicos entre los países pertenecientes a la zona monetaria. Por ello, se analizarán los resultados obtenidos del cálculo de las correlaciones dinámicas que existen entre los indicadores de confianza de los once países que se incorporaron al euro en 1999. La finalidad principal es poder identificar si los países analizados se ven afectados de la misma manera en las distintas fases de los ciclos económicos.

PURPOSE: The research aimed to verify the applicability of the SWOT / TOWS method to the selection of organizational health and safety management strategies in manufacturing companies, with a focus on the correlation approach in the conducted research., DESIGN/METHODOLOGY/APPROACH: Determining a strategy in the field of health and safety management in nine production companies. The methodology of selecting a strategy was based on SWOT/TOWS methodology assumption. The input data for the research was: characteristics of the company's operations, description of the context of the company's operation, requirements of interested parties along with their criticality based on the prepared hierarchy of occupational health and safety(OHS) requirements., FINDINGS: The use of the correlation approach in determining the OHS management strategy allowed for estimating the strength of the impact of individual factors on the selection of OHS management strategies in a manufacturing company, identification of interactions between factors in 8 crosstables, and above all, determination of OHS management strategies, i.e. for 7 companies of the strategy aggressive, for the others it is conservative and defensive., PRACTICAL IMPLICATIONS: Establishing a health and safety management strategy should be considered particularly important in the context of the challenges faced by enterprises resulting from the COVID-19 pandemic. As a result of the conducted research, it can be concluded that the SWOT / TOWS method enables the determination of an occupational health and safety management strategy., ORIGINALITY/VALUE: In the SWOT / TOWS method, attention should be paid to the dominant number of interactions and to the distribution of the weighted number of interactions. The article shows that there are correlations for which it is necessary to indicate an action strategy to a specific combination of factors, especially when the product of weights and the number of interactions indicates the priority of a strategy other than that adopted for the entire enterprise., This article was funded by Poznan University of Technology, Faculty of Engineering Management [project number: SBAD:0812/SBAD/4185]., peer-reviewed

PURPOSE: This study was carried out to analyze the structure of the aggregated network at the level of economic sectors and to reveal the central/peripheral sectors., DESIGN/METHODOLOGY/APPROACH: The study uses the method of complex networks, with the two-step procedure employed to construct the network of economic sectors. First, the MST approach is utilized based on the cross-correlation of 496 stock price returns of the S&P500 Index. Then, the network is aggregated at the level of economic sectors. In addition, to analyze the graph, the network theory, multi-dimensional scaling (MDS), and relative importance approaches are employed., FINDINGS: The results indicate that the sectoral network has a core/periphery structure. Based on the centrality measures, the ranking of sectors is provided. Of the 11 sectors, 3 are classified as central nodes, 4 as peripheral nodes, and the remaining 4 are classified as intermediate. In addition, the network configuration analysis demonstrates that the graph consists of two parts with a star-like structure, connected through the industrials sector., PRACTICAL IMPLICATIONS: An analysis of the cross-correlation network of aggregated assets at the level of economic sectors can be applied to ascertain the direction of stock price movements in the stock market. The division of sectors in the network into central and peripheral nodes has important implications for the management of an optimal portfolio of stocks., ORIGINALITY/VALUE: This study contributes to complex network theory and portfolio strategy design. A unique procedure is proposed to construct the network of economic sectors using the MST-based approach. Detection of the stock market network structure is vital for investors and regulators alike., peer-reviewed

PURPOSE: The paper presents the application of a test based on the inversions number distribution. The purpose of the work is to examine the stability of dependence measures between rates of return for financial instruments., DESIGN/METHODOLOGY/APPROACH: The issue plays a vital role in portfolio analysis because dependence measures are an essential factor affecting the portfolio risk of financial instruments. Direct application of the classic measure of dependence, such as Pearson's correlation coefficient, requires several additional assumptions regarding the existence of moments or the normality of the distribution of random variables analyzed. Therefore, the choice falls on the tau Kendall coefficient., FINDINGS: First of all, it is suitable for testing distributions for which variance does not exist (e.g., stable distributions). Secondly, we present a new way of testing the hypothesis about the equality of the Kendall coefficient over two separate periods. A miniature sample version was introduced in the article. An essential feature of the proposed test is the ability to calculate its power analytically., peer-reviewed

Microwave radiometers can be considerably affected by Radio Frequency Interference (RFI). These man-made interference conceal the underlying natural signal, preventing the retrieval of geophysical variables. Adopting on-board detection and mitigation techniques is a requirement to reduce the impact of RFI. Several families of RFI detection algorithms have been developed over the last years (e.g. [1]–[4]). In this work, a new detection technique is proposed and its performance analyzed. It is based in the distortion of the shape of the cross-correlation function at lags different from zero under the presence of RFI. Its performance is compared to other common RFI mitigation algorithms. Proposed methods' performance is found to surpass other common algorithms such as signal Kurtosis, while presenting some convenient properties for its practical application in correlation and synthetic aperture radiometers.

Aquesta tesi de final de grau forma part d'un projecte del Grup de Processament d'Imatge de la UPC enfocat a la detecció de llenguatge de signes utilitzant tecnologies relacionades amb deep learning. Aquest projecte ja consta amb una base de dades anomenada How2sign, que conté més de 83 hores de vídeos de traducció de llenguatge de signes. Aquesta base de dades conté anotacions textuals alineades a una càmera RGB frontal. Les mateixes escenes també són capturades per una RGB lateral i una RGB-D frontal. Aquestes tres càmeres no estan sincronitzades, amb la qual cosa és necessari alinear els segments anotats de la RGB frontal amb les altres. En aquesta tesi s'explora una primera solució basada en la correlació creuada. El nostre treball consisteix a processar els punts de les coordenades de les articulacions de l'subjecte que apareix en els vídeos, no des del punt de vista de processament d'imatge o vídeo basat en píxels. La primera part d'aquesta tesi és investigar les propietats de la funció de correlació creuada mitjançant la localització de segments curts de vídeo d'una gravació llarga basada en l'extracció automàtica de les poses en 2D. Els experiments també estudien l'impacte d'afegir soroll. La segona aplica la correlació creuada per intentar alinear dos videos amb el mateix contingut, però gravats amb diferents càmeres des de diferents punts de vista. Esta tesis de final de grado forma parte de un proyecto del Grupo de Procesado de Imagen de la UPC enfocado a la detección de lenguaje de signos utilizando tecnologías relacionadas con deep learning. Este proyecto ya consta con una base de datos llamada How2sign, que contiene más de 83 horas de videos de traducción de lenguaje de signos. Esta base de datos contiene anotaciones textuales alineadas a una cámara RGB frontal. Las mismas escenas también son capturadas por una RGB lateral y una RGB-D frontal. Estas tres cámaras no están sincronizadas, con lo cual es necesario alinear los segmentos anotados de la RGB frontal con las demás. En esta tesis se explora una primera solucion basada en la correlación cruzada. Nuestro trabajo consiste en procesar los puntos de las coordenadas de las articulaciones del sujeto que aparece en los videos, no desde el punto de vista de procesado de imagen o video basado en píxeles. La primera parte de esta tesis es investigar las propiedades de la función de correlación cruzada mediante la localización de segmentos cortos de vídeo de una grabación larga basada en la extracción automática de las poses en 2D. Los experimentos también estudian el impacto de añadir ruido. La segunda aplica la correlación cruzada para intentar alinear dos videos con el mismo contenido, pero grabados con distintas cámaras desde distintos puntos de vista. This thesis degree is part of a project from the Image Group at UPC that is focused on sign language translation using deep learning technologies. This thesis builds on top of an existing database called How2Sign, that contains more than 83 hours of sign language translation videos. This database has some textual annotations aligned to a front RGB camera. The same scenes are also captured by a side RGB and a front RGB-D cameras. These three cameras are not synchronized, so it is necessary to align the segments annotated on the RGB front camera to the other cameras. This thesis explores a solution based on the cross correlation operator. Our work is to process the coordinates of the joints of the subject that appears in the videos, not from the point of view of image or video processing based on pixels. The first part if this thesis is to investigate the properties of the cross-correlation function by locating short video segments of a long recording based on automatically extracted 2D human poses. The experiments studied the impact of adding noise. The second part applied the cross-correlation to try to align two videos with the same scene, but recorded with different cameras from different points of view.