Aplicação de classificadores Bayesianos e regressão logística na análise de desempenho dos alunos de graduação

Este trabalho minera as informações coletadas no processo de vestibular entre 2009 e 2012 para o curso de graduação de administração de empresas da FGV-EAESP, para estimar classificadores capazes de calcular a probabilidade de um novo aluno ter bom desempenho. O processo de KDD (Knowledge Discovery in Database) desenvolvido por Fayyad et al. (1996a) é a base da metodologia adotada e os classificadores serão estimados utilizando duas ferramentas matemáticas. A primeira é a regressão logística, muito usada por instituições financeiras para avaliar se um cliente será capaz de honrar com seus pagamentos e a segunda é a rede Bayesiana, proveniente do campo de inteligência artificial. Este estudo mostre que os dois modelos possuem o mesmo poder discriminatório, gerando resultados semelhantes. Além disso, as informações que influenciam a probabilidade de o aluno ter bom desempenho são a sua idade no ano de ingresso, a quantidade de vezes que ele prestou vestibular da FGV/EAESP antes de ser aprovado, a região do Brasil de onde é proveniente e as notas das provas de matemática fase 01 e fase 02, inglês, ciências humanas e redação. Aparentemente o grau de formação dos pais e o grau de decisão do aluno em estudar na FGV/EAESP não influenciam nessa probabilidade.

Delimiting genetic units in Neotropical toads under incomplete lineage sorting and hybridization

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Ambiente independente de idioma para suporte a identificação de tuplas duplicadas por meio da similaridade fonética e numérica: otimização de algoritmo baseado em multithreading

Pós-graduação em Ciência da Computação - IBILCE

Identificação e estimação de ruído em redes DSL: uma abordagem baseada em inteligência computacional

Este trabalho propõe a utilização de técnicas de inteligência computacional objetivando identificar e estimar a potencia de ruídos em redes Digital Subscriber Line ou Linhas do Assinante Digital (DSL) em tempo real. Uma metodologia baseada no Knowledge Discovery in Databases ou Descobrimento de Conhecimento em Bases de Dados (KDD) para detecção e estimação de ruídos em tempo real, foi utilizada. KDD é aplicado para selecionar, pré-processar e transformar os dados antes da etapa de aplicação dos algoritmos na etapa de mineração de dados. Para identificação dos ruídos o algoritmo tradicional backpropagation baseado em Redes Neurais Artificiais (RNA) é aplicado objetivando identificar o tipo de ruído em predominância durante a coleta das informações do modem do usuário e da central. Enquanto, para estimação o algoritmo de regressão linear e o algoritmo híbrido composto por Fuzzy e regressão linear foram aplicados para estimar a potência em Watts de ruído crosstalk ou diafonia na rede. Os resultados alcançados demonstram que a utilização de algoritmos de inteligência computacional como a RNA são promissores para identificação de ruídos em redes DSL, e que algoritmos como de regressão linear e Fuzzy com regressão linear (FRL) são promissores para a estimação de ruídos em redes DSL.

Imputação de dados baseado em otimização por enxame de partículas considerando os principais mecanismos de ausência de dados

Durante o processo de extração do conhecimento em bases de dados, alguns problemas podem ser encontrados como por exemplo, a ausência de determinada instância de um atributo. A ocorrência de tal problemática pode causar efeitos danosos nos resultados finais do processo, pois afeta diretamente a qualidade dos dados a ser submetido a um algoritmo de aprendizado de máquina. Na literatura, diversas propostas são apresentadas a fim de contornar tal dano, dentre eles está a de imputação de dados, a qual estima um valor plausível para substituir o ausente. Seguindo essa área de solução para o problema de valores ausentes, diversos trabalhos foram analisados e algumas observações foram realizadas como, a pouca utilização de bases sintéticas que simulem os principais mecanismos de ausência de dados e uma recente tendência a utilização de algoritmos bio-inspirados como tratamento do problema. Com base nesse cenário, esta dissertação apresenta um método de imputação de dados baseado em otimização por enxame de partículas, pouco explorado na área, e o aplica para o tratamento de bases sinteticamente geradas, as quais consideram os principais mecanismos de ausência de dados, MAR, MCAR e NMAR. Os resultados obtidos ao comprar diferentes configurações do método à outros dois conhecidos na área (KNNImpute e SVMImpute) são promissores para sua utilização na área de tratamento de valores ausentes uma vez que alcançou os melhores valores na maioria dos experimentos realizados.

Otimização do processo de aprendizagem da estrutura gráfica de Redes Bayesianas em BigData

A automação na gestão e análise de dados tem sido um fator crucial para as empresas que necessitam de soluções eficientes em um mundo corporativo cada vez mais competitivo. A explosão do volume de informações, que vem se mantendo crescente nos últimos anos, tem exigido cada vez mais empenho em buscar estratégias para gerenciar e, principalmente, extrair informações estratégicas valiosas a partir do uso de algoritmos de Mineração de Dados, que comumente necessitam realizar buscas exaustivas na base de dados a fim de obter estatísticas que solucionem ou otimizem os parâmetros do modelo de extração do conhecimento utilizado; processo que requer computação intensiva para a execução de cálculos e acesso frequente à base de dados. Dada a eficiência no tratamento de incerteza, Redes Bayesianas têm sido amplamente utilizadas neste processo, entretanto, à medida que o volume de dados (registros e/ou atributos) aumenta, torna-se ainda mais custoso e demorado extrair informações relevantes em uma base de conhecimento. O foco deste trabalho é propor uma nova abordagem para otimização do aprendizado da estrutura da Rede Bayesiana no contexto de BigData, por meio do uso do processo de MapReduce, com vista na melhora do tempo de processamento. Para tanto, foi gerada uma nova metodologia que inclui a criação de uma Base de Dados Intermediária contendo todas as probabilidades necessárias para a realização dos cálculos da estrutura da rede. Por meio das análises apresentadas neste estudo, mostra-se que a combinação da metodologia proposta com o processo de MapReduce é uma boa alternativa para resolver o problema de escalabilidade nas etapas de busca em frequência do algoritmo K2 e, consequentemente, reduzir o tempo de resposta na geração da rede.

Collaborative fuzzy clustering algorithms: some refinements and design guidelines

There are some variants of the widely used Fuzzy C-Means (FCM) algorithm that support clustering data distributed across different sites. Those methods have been studied under different names, like collaborative and parallel fuzzy clustering. In this study, we offer some augmentation of the two FCM-based clustering algorithms used to cluster distributed data by arriving at some constructive ways of determining essential parameters of the algorithms (including the number of clusters) and forming a set of systematically structured guidelines such as a selection of the specific algorithm depending on the nature of the data environment and the assumptions being made about the number of clusters. A thorough complexity analysis, including space, time, and communication aspects, is reported. A series of detailed numeric experiments is used to illustrate the main ideas discussed in the study.

Body ares networks for human motor assessment

Healthcare, Human Computer Interfaces (HCI), Security and Biometry are the most promising application scenario directly involved in the Body Area Networks (BANs) evolution. Both wearable devices and sensors directly integrated in garments envision a word in which each of us is supervised by an invisible assistant monitoring our health and daily-life activities. New opportunities are enabled because improvements in sensors miniaturization and transmission efficiency of the wireless protocols, that achieved the integration of high computational power aboard independent, energy-autonomous, small form factor devices. Application’s purposes are various: (I) data collection to achieve off-line knowledge discovery; (II) user notification of his/her activities or in case a danger occurs; (III) biofeedback rehabilitation; (IV) remote alarm activation in case the subject need assistance; (V) introduction of a more natural interaction with the surrounding computerized environment; (VI) users identification by physiological or behavioral characteristics. Telemedicine and mHealth [1] are two of the leading concepts directly related to healthcare. The capability to borne unobtrusiveness objects supports users’ autonomy. A new sense of freedom is shown to the user, not only supported by a psychological help but a real safety improvement. Furthermore, medical community aims the introduction of new devices to innovate patient treatments. In particular, the extension of the ambulatory analysis in the real life scenario by proving continuous acquisition. The wide diffusion of emerging wellness portable equipment extended the usability of wearable devices also for fitness and training by monitoring user performance on the working task. The learning of the right execution techniques related to work, sport, music can be supported by an electronic trainer furnishing the adequate aid. HCIs made real the concept of Ubiquitous, Pervasive Computing and Calm Technology introduced in the 1988 by Marc Weiser and John Seeley Brown. They promotes the creation of pervasive environments, enhancing the human experience. Context aware, adaptive and proactive environments serve and help people by becoming sensitive and reactive to their presence, since electronics is ubiquitous and deployed everywhere. In this thesis we pay attention to the integration of all the aspects involved in a BAN development. Starting from the choice of sensors we design the node, configure the radio network, implement real-time data analysis and provide a feedback to the user. We present algorithms to be implemented in wearable assistant for posture and gait analysis and to provide assistance on different walking conditions, preventing falls. Our aim, expressed by the idea to contribute at the development of a non proprietary solutions, driven us to integrate commercial and standard solutions in our devices. We use sensors available on the market and avoided to design specialized sensors in ASIC technologies. We employ standard radio protocol and open source projects when it was achieved. The specific contributions of the PhD research activities are presented and discussed in the following. • We have designed and build several wireless sensor node providing both sensing and actuator capability making the focus on the flexibility, small form factor and low power consumption. The key idea was to develop a simple and general purpose architecture for rapid analysis, prototyping and deployment of BAN solutions. Two different sensing units are integrated: kinematic (3D accelerometer and 3D gyroscopes) and kinetic (foot-floor contact pressure forces). Two kind of feedbacks were implemented: audio and vibrotactile. • Since the system built is a suitable platform for testing and measuring the features and the constraints of a sensor network (radio communication, network protocols, power consumption and autonomy), we made a comparison between Bluetooth and ZigBee performance in terms of throughput and energy efficiency. Test in the field evaluate the usability in the fall detection scenario. • To prove the flexibility of the architecture designed, we have implemented a wearable system for human posture rehabilitation. The application was developed in conjunction with biomedical engineers who provided the audio-algorithms to furnish a biofeedback to the user about his/her stability. • We explored off-line gait analysis of collected data, developing an algorithm to detect foot inclination in the sagittal plane, during walk. • In collaboration with the Wearable Lab – ETH, Zurich, we developed an algorithm to monitor the user during several walking condition where the user carry a load. The remainder of the thesis is organized as follows. Chapter I gives an overview about Body Area Networks (BANs), illustrating the relevant features of this technology and the key challenges still open. It concludes with a short list of the real solutions and prototypes proposed by academic research and manufacturers. The domain of the posture and gait analysis, the methodologies, and the technologies used to provide real-time feedback on detected events, are illustrated in Chapter II. The Chapter III and IV, respectively, shown BANs developed with the purpose to detect fall and monitor the gait taking advantage by two inertial measurement unit and baropodometric insoles. Chapter V reports an audio-biofeedback system to improve balance on the information provided by the use centre of mass. A walking assistant based on the KNN classifier to detect walking alteration on load carriage, is described in Chapter VI.

IDENTIFICATION OF GENES CONTROLLING BIOLOGICAL PROCESSES AND PATHWAYS THROUGH STATISTICAL ANALYSIS AND NETWORK RECONSTRUCTION

The developmental processes and functions of an organism are controlled by the genes and the proteins that are derived from these genes. The identification of key genes and the reconstruction of gene networks can provide a model to help us understand the regulatory mechanisms for the initiation and progression of biological processes or functional abnormalities (e.g. diseases) in living organisms. In this dissertation, I have developed statistical methods to identify the genes and transcription factors (TFs) involved in biological processes, constructed their regulatory networks, and also evaluated some existing association methods to find robust methods for coexpression analyses. Two kinds of data sets were used for this work: genotype data and gene expression microarray data. On the basis of these data sets, this dissertation has two major parts, together forming six chapters. The first part deals with developing association methods for rare variants using genotype data (chapter 4 and 5). The second part deals with developing and/or evaluating statistical methods to identify genes and TFs involved in biological processes, and construction of their regulatory networks using gene expression data (chapter 2, 3, and 6). For the first part, I have developed two methods to find the groupwise association of rare variants with given diseases or traits. The first method is based on kernel machine learning and can be applied to both quantitative as well as qualitative traits. Simulation results showed that the proposed method has improved power over the existing weighted sum method (WS) in most settings. The second method uses multiple phenotypes to select a few top significant genes. It then finds the association of each gene with each phenotype while controlling the population stratification by adjusting the data for ancestry using principal components. This method was applied to GAW 17 data and was able to find several disease risk genes. For the second part, I have worked on three problems. First problem involved evaluation of eight gene association methods. A very comprehensive comparison of these methods with further analysis clearly demonstrates the distinct and common performance of these eight gene association methods. For the second problem, an algorithm named the bottom-up graphical Gaussian model was developed to identify the TFs that regulate pathway genes and reconstruct their hierarchical regulatory networks. This algorithm has produced very significant results and it is the first report to produce such hierarchical networks for these pathways. The third problem dealt with developing another algorithm called the top-down graphical Gaussian model that identifies the network governed by a specific TF. The network produced by the algorithm is proven to be of very high accuracy.

Knowledge transfer in software-maintenance offshore outsourcing

Software-maintenance offshore outsourcing (SMOO) projects have been plagued by tedious knowledge transfer during the service transition to the vendor. Vendor engineers risk being over-strained by the high amounts of novel information, resulting in extra costs that may erode the business case behind offshoring. Although stakeholders may desire to avoid these extra costs by implementing appropriate knowledge transfer practices, little is known on how effective knowledge transfer can be designed and managed in light of the high cognitive loads in SMOO transitions. The dissertation at hand addresses this research gap by presenting and integrating four studies. The studies draw on cognitive load theory, attributional theory, and control theory and they apply qualitative, quantitative, and simulation methods to qualitative data from eight in-depth longitudinal cases. The results suggest that the choice of appropriate learning tasks may be more central to knowledge transfer than the amount of information shared with vendor engineers. Moreover, because vendor staff may not be able to and not dare to effectively self-manage learn-ing tasks during early transition, client-driven controls may be initially required and subsequently faded out. Collectively, the results call for people-based rather than codification-based knowledge management strategies in at least moderately specific and complex software environments.

Combining data mining and ontology engineering to enrich ontologies and linked data

In this position paper, we claim that the need for time consuming data preparation and result interpretation tasks in knowledge discovery, as well as for costly expert consultation and consensus building activities required for ontology building can be reduced through exploiting the interplay of data mining and ontology engineering. The aim is to obtain in a semi-automatic way new knowledge from distributed data sources that can be used for inference and reasoning, as well as to guide the extraction of further knowledge from these data sources. The proposed approach is based on the creation of a novel knowledge discovery method relying on the combination, through an iterative ?feedbackloop?, of (a) data mining techniques to make emerge implicit models from data and (b) pattern-based ontology engineering to capture these models in reusable, conceptual and inferable artefacts.

Nanoinformatics knowledge infrastructures: bringing efficient information management to nanomedical research

Nanotechnology represents an area of particular promise and significant opportunity across multiple scientific disciplines. Ongoing nanotechnology research ranges from the characterization of nanoparticles and nanomaterials to the analysis and processing of experimental data seeking correlations between nanoparticles and their functionalities and side effects. Due to their special properties, nanoparticles are suitable for cellular-level diagnostics and therapy, offering numerous applications in medicine, e.g. development of biomedical devices, tissue repair, drug delivery systems and biosensors. In nanomedicine, recent studies are producing large amounts of structural and property data, highlighting the role for computational approaches in information management. While in vitro and in vivo assays are expensive, the cost of computing is falling. Furthermore, improvements in the accuracy of computational methods (e.g. data mining, knowledge discovery, modeling and simulation) have enabled effective tools to automate the extraction, management and storage of these vast data volumes. Since this information is widely distributed, one major issue is how to locate and access data where it resides (which also poses data-sharing limitations). The novel discipline of nanoinformatics addresses the information challenges related to nanotechnology research. In this paper, we summarize the needs and challenges in the field and present an overview of extant initiatives and efforts.

Estudio de patrones de interacción entre los estudiantes y la Plataforma de Tele-Enseñanza en la UPM

Vivimos en una sociedad en la que la información ha adquirido una vital importancia. El uso de Internet y el desarrollo de nuevos sistemas de la información han generado un ferviente interés tanto de empresas como de instituciones en la búsqueda de nuevos patrones que les proporcione la clave del éxito. La Analítica de Negocio reúne un conjunto de herramientas, estrategias y técnicas orientadas a la explotación de la información con el objetivo de crear conocimiento útil dentro de un marco de trabajo y facilitar la optimización de los recursos tanto de empresas como de instituciones. El presente proyecto se enmarca en lo que se conoce como Gestión Educativa. Se aplicará una arquitectura y modelo de trabajo similar a lo que se ha venido haciendo en los últimos años en el entorno empresarial con la Inteligencia de Negocio. Con esta variante, se pretende mejorar la calidad de la enseñanza, agilizar las decisiones dentro de la institución académica, fortalecer las capacidades del cuerpo docente y en definitiva favorecer el aprendizaje del alumnado. Para lograr el objetivo se ha decidido seguir las etapas del Knowledge Discovery in Databases (KDD), una de las metodologías más conocidas dentro de la Inteligencia de Negocio, que describe el procedimiento que va desde la selección de la información y su carga en sistemas de almacenamiento, hasta la aplicación de técnicas de minería de datos para la obtención nuevo conocimiento. Los estudios se realizan a partir de la información de la activad de los usuarios dentro la plataforma de Tele-Enseñanza de la Universidad Politécnica de Madrid (Moodle). Se desarrollan trabajos de extracción y preprocesado de la base de datos en crudo y se aplican técnicas de minería de datos. En la aplicación de técnicas de minería de datos, uno de los factores más importantes a tener en cuenta es el tipo de información que se va a tratar. Por este motivo, se trabaja con la Minería de Datos Educativa, en inglés, Educational Data Mining (EDM) que consiste en la aplicación de técnicas de minería optimizadas para la información que se genera en entornos educativos. Dentro de las posibilidades que ofrece el EDM, se ha decidido centrar los estudios en lo que se conoce como analítica predictiva. El objetivo fundamental es conocer la influencia que tienen las interacciones alumno-plataforma en las calificaciones finales y descubrir nuevas reglas que describan comportamientos que faciliten al profesorado discriminar si un estudiante va a aprobar o suspender la asignatura, de tal forma que se puedan tomar medidas que mejoren su rendimiento. Toda la información tratada en el presente proyecto ha sido previamente anonimizada para evitar cualquier tipo de intromisión que atente contra la privacidad de los elementos participantes en el estudio. ABSTRACT. We live in a society dominated by data. The use of the Internet accompanied by developments in information systems has generated a sustained interest among companies and institutions to discover new patterns to succeed in their business ventures. Business Analytics (BA) combines tools, strategies and techniques focused on exploiting the available information, to optimize resources and create useful insight. The current project is framed under Educational Management. A Business Intelligence (BI) architecture and business models taught up to date will be applied with the aim to accelerate the decision-making in academic institutions, strengthen teacher´s skills and ultimately improve the quality of teaching and learning. The best way to achieve this is to follow the Knowledge Discovery in Databases (KDD), one of the best-known methodologies in B.I. This process describes data preparation, selection, and cleansing through to the application of purely Data Mining Techniques in order to incorporate prior knowledge on data sets and interpret accurate solutions from the observed results. The studies will be performed using the information extracted from the Universidad Politécnica de Madrid Learning Management System (LMS), Moodle. The stored data is based on the user-platform interaction. The raw data will be extracted and pre-processed and afterwards, Data Mining Techniques will be applied. One of the crucial factors in the application of Data Mining Techniques is the kind of information that will be processed. For this reason, a new Data Mining perspective will be taken, called Educational Data Mining (EDM). EDM consists of the application of Data Mining Techniques but optimized for the raw data generated by the educational environment. Within EDM, we have decided to drive our research on what is called Predictive Analysis. The main purpose is to understand the influence of the user-platform interactions in the final grades of students and discover new patterns that explain their behaviours. This could allow teachers to intervene ahead of a student passing or failing, in such a way an action could be taken to improve the student performance. All the information processed has been previously anonymized to avoid the invasion of privacy.

Towards a reverse engineering approach for guiding user in applying data mining

Comunicación presentada en las XVI Jornadas de Ingeniería del Software y Bases de Datos, JISBD 2011, A Coruña, 5-7 septiembre 2011.

Origins and early development of human body knowledge

As a knowable object, the human body is highly complex. Evidence from several converging lines of research, including psychological studies, neuroimaging and clinical neuropsychology, indicates that human body knowledge is widely distributed in the adult brain, and is instantiated in at least three partially independent levels of representation. Sensori-motor body knowledge is responsible for on-line control and movement of one's own body and may also contribute to the perception of others' moving bodies; visuo-spatial body knowledge specifies detailed structural descriptions of the spatial attributes of the human body; and lexical-semantic body knowledge contains language-based knowledge about the human body. In the first chapter of this Monograph, we outline the evidence for these three hypothesized levels of human body knowledge, then review relevant literature on infants' and young children's human body knowledge in terms of the three-level framework. In Chapters II and III, we report two complimentary series of studies that specifically investigate the emergence of visuospatial body knowledge in infancy. Our technique is to compare infants' responses to typical and scrambled human bodies, in order to evaluate when and how infants acquire knowledge about the canonical spatial layout of the human body. Data from a series of visual habituation studies indicate that infants first discriminate scrambled from typical human body pictures at 15 to 18 months of age. Data from object examination studies similarly indicate that infants are sensitive to violations of three-dimensional human body stimuli starting at 15-18 months of age. The overall pattern of data supports several conclusions about the early development of human body knowledge: (a) detailed visuo-spatial knowledge about the human body is first evident in the second year of life, (b) visuo-spatial knowledge of human faces and human bodies are at least partially independent in infancy and (c) infants' initial visuo-spatial human body representations appear to be highly schematic, becoming more detailed and specific with development. In the final chapter, we explore these conclusions and discuss how levels of body knowledge may interact in early development.