Submitted by Clebson Anjos (clebson.leandro54@gmail.com) on 2016-02-15T18:37:51Z No. of bitstreams: 1 arquivototal.pdf: 2981698 bytes, checksum: 6dfa47590c870db030e7c1cbea499120 (MD5) / Made available in DSpace on 2016-02-15T18:37:51Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 2981698 bytes, checksum: 6dfa47590c870db030e7c1cbea499120 (MD5) Previous issue date: 2015-05-29 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Through the use of data mining techniques, more usually the classification algorithms, it is possible to implement predictive models that are able to early identify a student in risk of dropout. Several studies used data obtained from a Virtual Learning Environment (VLE) to implement predictive performance models in a discipline of a course. However, any study was carried out aimed at developing a model for dropout prediction, to distance graduation courses of longer duration, which integrates works that carry out performance prediction based on a VLE, allowing an early prediction during the first semester and throughout the others semesters. Thus, this work proposes a dropout identification approach for distance graduation courses that use the Rule-Based Classification technique to firstly identify the disciplines and grades limits that have higher influence on dropout, so that the predictive models for performance in a VLE can be used regarding the dropout detection of students along the whole distance graduation course. Experiments were carried out using four rulebased classification algorithms: JRip, OneR, PART and Ridor. Considering the use of this temporal approach, it was possible to prove the advantages of this approach, once better accuracies were obtained along the semesters and important rules were discovered to early identify students in risk of dropout. Among the applied algorithms, JRip and PART obtained the best predictive results with average accuracy of 81% at the end of first semester. Furthermore, considering our proposed partition methodology, where attributes of the predictive models are incrementally applied, it was possible to discovery rules potentially useful to dropout prevention. / Com a utilização de técnicas de mineração de dados, mais comumente os algoritmos de Classificação, pode-se construir modelos preditivos capazes de identificar precocemente um estudante com risco de evasão. Diversos estudos utilizaram dados obtidos de um Ambiente Virtual de Aprendizagem (AVA) para a construção de modelos preditivos de desempenho em uma disciplina de um curso. Porém, nenhum estudo foi realizado com o objetivo de desenvolver um modelo de predição de evasão, para um curso de graduação a distância de maior duração, que integre trabalhos que fazem a predição de desempenho a partir de um AVA, possibilitando uma predição da evasão antecipada durante o primeiro semestre e ao longo dos demais semestres. Assim, este trabalho propõe uma abordagem de identificação de evasão em um curso de graduação a distância a partir da utilização da técnica de classificação baseada em regras para, primeiramente, identificar as disciplinas e os limites de notas que mais influenciam na evasão para que os modelos preditivos de desempenhos em um AVA possam ser utilizados para a predição da evasão de um aluno com risco de evasão ao longo de todo o curso de graduação a distância. Foram realizados experimentos com quatro algoritmos de classificação baseados em regras: o JRip, o OneR, o PART e o Ridor. A partir da utilização da abordagem temporal proposta foi possível comprovar sua vantagem, uma vez que foram obtidos melhores desempenhos preditivos ao longo dos semestres e foram descobertas importantes regras para a identificação precoce de um estudante com risco de evasão. Entre os algoritmos estudados, JRip e PART obtiveram os melhores desempenhos preditivos com acurácia média de 81% ao final do primeiro semestre. A partir da metodologia proposta de partições, na qual os atributos dos modelos preditivos são aplicados de forma incremental, foi possível a descoberta de regras potencialmente úteis para prevenir a evasão.

L'extraction de règles d'association et de bi-clusters sont deux techniques de fouille de données complémentaires majeures, notamment pour l'intégration de connaissances. Ces techniques sont utilisées dans de nombreux domaines, mais aucune approche permettant de les unifier n'a été proposée. Hors, réaliser ces extractions indépendamment pose les problèmes des ressources nécessaires (mémoire, temps d'exécution et accès aux données) et de l'unification des résultats. Nous proposons une approche originale pour extraire différentes catégories de modèles de connaissances tout en utilisant un minimum de ressources. Cette approche est basée sur la théorie des ensembles fermés et utilise une nouvelle structure de données pour extraire des représentations conceptuelles minimales de règles d'association, bi-clusters et règles de classification. Ces modèles étendent les règles d'association et de classification et les bi-clusters classiques, les listes d'objets supportant chaque modèle et les relations hiérarchiques entre modèles étant également extraits. Cette approche a été appliquée pour l'analyse de données d'interaction protéomiques entre le virus VIH-1 et l'homme. L'analyse de ces interactions entre espèces est un défi majeur récent en bio-informatique. Plusieurs bases de données intégrant des informations hétérogènes sur les interactions et des connaissances biologiques sur les protéines ont été construites. Les résultats expérimentaux montrent que l'approche proposée peut traiter efficacement ces bases de données et que les modèles conceptuels extraits peuvent aider à la compréhension et à l'analyse de la nature des relations entre les protéines interagissant. / Knowledge pattern extraction is one of the major topics in the data mining and background knowledge integration domains. Out of several data mining techniques, association rule mining and bi-clustering are two major complementary tasks for these topics. These tasks gained much importance in many domains in recent years. However, no approach was proposed to perform them in one process. This poses the problems of resources required (memory, execution times and data accesses) to perform independent extractions and of the unification of the different results. We propose an original approach for extracting different categories of knowledge patterns while using minimum resources. This approach is based on the frequent closed patterns theoretical framework and uses a novel suffix-tree based data structure to extract conceptual minimal representations of association rules, bi-clusters and classification rules. These patterns extend the classical frameworks of association and classification rules, and bi-clusters as data objects supporting each pattern and hierarchical relationships between patterns are also extracted. This approach was applied to the analysis of HIV-1 and human protein-protein interaction data. Analyzing such inter-species protein interactions is a recent major challenge in computational biology. Databases integrating heterogeneous interaction information and biological background knowledge on proteins have been constructed. Experimental results show that the proposed approach can efficiently process these databases and that extracted conceptual patterns can help the understanding and analysis of the nature of relationships between interacting proteins.