There are common pitfalls and neglected areas when using clustering approaches to solve educational problems. A clustering algorithm is often used without the choice being justified. Few comparisons between a selected algorithm and a competing algorithm are presented, and results are presented without validation. Lastly, few studies fully utilize data provided in an educational environment to evaluate their findings. In response to these problems, this thesis describes a rigorous study comparing two clustering algorithms in the context of an educational digital library service, called the Instructional Architect. First, a detailed description of the chosen clustering algorithm, namely, latent class analysis (LCA), is presented. Second, three kinds of preprocessed data are separately applied to both the selected algorithm and a competing algorithm, namely, K-means algorithm. Third, a series of comprehensive evaluations on four aspects of each clustering result, i.e., intra-cluster and inter-cluster distances, Davies-Bouldin index, users' demographic profile, and cluster evolution, are conducted to compare the clustering results of LCA and K-means algorithms. Evaluation results show that LCA outperforms K-means in producing consistent clustering results at different settings, finding compact clusters, and finding connections between users' teaching experience and their effectiveness in using the IA. The implication, contributions, and limitation of this research are discussed.

Data Structures and Algorithms (DSA) courses are critical for any computer science curriculum. DSA courses emphasize concepts related to procedural dynamics and Algorithm Analysis (AA). These concepts are hard for students to grasp when conveyed using traditional textbook material relying on text and static images. Algorithm Visualizations (AVs) emerged as a technique for conveying DSA concepts using interactive visual representations. Historically, AVs have dealt with portraying algorithm dynamics, and the AV developer community has decades of successful experience with this. But there exist few visualizations to present algorithm analysis concepts. This content is typically still conveyed using text and static images. We have devised an approach that we term Algorithm Analysis Visualizations (AAVs), capable of conveying AA concepts visually. In AAVs, analysis is presented as a series of slides where each statement of the explanation is connected to visuals that support the sentence. We developed a pool of AAVs targeting the basic concepts of AA. We also developed AAVs for basic sorting algorithms, providing a concrete depiction about how the running time analysis of these algorithms can be calculated. To evaluate AAVs, we conducted a quasi-experiment across two offerings of CS3114 at Virginia Tech. By analyzing OpenDSA student interaction logs, we found that intervention group students spent significantly more time viewing the material as compared to control group students who used traditional textual content. Intervention group students gave positive feedback regarding the usefulness of AAVs to help them understand the AA concepts presented in the course. In addition, intervention group students demonstrated better performance than control group students on the AA part of the final exam. The final exam taken by both the control and intervention groups was based on a pilot version of the Algorithm Analysis Concept Inventory (AACI) that was developed to target fundamental AA concepts and probe students' misconceptions about these concepts. The pilot AACI was developed using a Delphi process involving a group of DSA instructors, and was shown to be a valid and reliable instrument to gauge students' understanding of the basic AA topics. / Ph. D.

Sistemas educacionais possuem diversas funcionalidades capazes de apoiar a interação entre alunos e professores de maneira dinâmica, síncrona e assíncrona. Uma das formas de monitorar a eficácia do processo educacional e por meio da utilização dos dados armazenados nesses sistemas como fonte de informação. Pesquisas em Learning Analytics, Academic Analytics e Mineração de Dados Educacionais, buscam explorar os dados de sistemas educacionais utilizando processamento analítico e técnicas de mineração de dados. No entanto, há uma serie de fatores que dificultam a gestão eficiente do processo educacional a partir dos dados de sistemas educacionais. A transformação de dados provenientes de diferentes tipos de sistemas educacionais, como Sistemas de Gestão de Aprendizagem e Sistemas Acadêmicos, e uma tarefa complexa devido a natureza heterogênea dos dados. Dados provenientes desses sistemas podem ser analisados considerando diferentes stakeholders, sob varias perspectivas e níveis de granularidade. Neste cenário, um modelo de referência para a descoberta de conhecimento a partir de dados de sistemas educacionais, denominado Modelo de Referência de Dados Educacionais (EDRM), foi desenvolvido neste trabalho. O EDRM e um modelo dimensional no formato star schema, estruturado em um Data Warehouse, projetado para ser uma fonte única de dados integrados e correlacionados voltada a tomada de decisão. Assim, e possível armazenar dados de diversas fontes, combina-los e, por fim, realizar analises que levem as instituições a desenvolver uma melhor compreensão, rastrear tendências e descobrir lacunas e ineficiências acerca do processo educacional. Neste trabalho, o EDRM foi validado por meio de um estudo de caso, utilizando bases de dados reais coletadas de diferentes sistemas educacionais. Os resultados mostram que o EDRM e eficiente em tarefas com diferentes objetivos, utilizando processamento analítico e mineração de dados. / Educational systems support dynamic, synchronous and asynchronous interaction between students and educators. Researches in Learning Analytics, Academic Analytics and Educational Data Mining explore data from educational systems for knowledge discovery through analytical processing, statistical analysis and data mining. However, there are some factors that hinder an efficient management of the educational process. The transformation of data from different kinds of educational system, as Learning Management Systems and Student Information Systems, can be even more difficult due to data heterogeneity. Data from these systems can be analyzed considering different stakeholders, under different perspectives and under different granularities. Motivated by this scenario, in this work we propose Modelo de Referência de Dados Educacionais (EDRM), a reference data model for knowledge discovery in data from educational systems. EDRM is an analytical model structured under a Data Warehouse architecture following a multidimensional data model. EDRM is projected for being an resource of integrated and correlated data focused in decision taking in the educational process. EDRM was developed considering a deep analysis of data and functionalities from different educational systems. In this sense, data from different kinds of systems and sources can be used unified, integrated and consistently. This allows institutions to better comprehend their data, as well as discover patterns, gaps and inefficiencies about their educational process. In this work, EDRM was validated in a case study using real-world databases from different educational systems. The results indicate that EDRM is efficient in tasks with different objectives, using Learning Analytics and Educational Data Mining techniques, and analyzing different perspectives.

The goal of Intelligent Tutoring systems (ITSs) is to engage the students in sustained reasoning activity and to interact with students based on a deep understanding of student behavior. In order to understand student behavior, ITSs rely on student modeling methods to observes student actions in the tutor and creates a quantitative representation of student knowledge, interests, affective states. Good student models are going to effectively help ITSs customize instructions, engage student's interest and then promote learning. Thus, the work of building ITSs and advancing student modeling should be considered as two interconnected components of one system rather than two separate topics. In this work, we utilized the theoretical support of a well-known learning science theory, the spacing effect, to guide the development of an ITS, called Automatic Reassessment and Relearning System (ARRS). ARRS not only validated the effectiveness of spacing effect, but it also served as a testing field which allowed us to find out new approaches to improve student learning by conducting large-scale randomized controlled trials (RCTs). The rich data set we gathered from ARRS has advanced our understanding of robust learning and helped us build student models with advanced data mining methods. At the end, we designed a set of API that supports the development of ARRS in next generation ASSISTments platform and adopted deep learning algorithms to further improve retention performance prediction. We believe our work is a successful example of combining theory and practice to advance science and address real- world problems.

In this dissertation, several approaches I have taken to build upon the student learning model are described. There are two focuses of this dissertation. The first focus is on improving the accuracy with which future student knowledge and performance can be predicted by individualizing the model to each student. The second focus is to predict how different educational content and tutorial strategies will influence student learning. The two focuses are complimentary but are approached from slightly different directions. I have found that Bayesian Networks, based on belief propagation, are strong at achieving the goals of both focuses. In prediction, they excel at capturing the temporal nature of data produced where student knowledge is changing over time. This concept of state change over time is very difficult to capture with classical machine learning approaches. Interpretability is also hard to come by with classical machine learning approaches; however, it is one of the strengths of Bayesian models and aids in studying the direct influence of various factors on learning. The domain in which these models are being studied is the domain of computer tutoring systems, software which uses artificial intelligence to enhance computer based tutorial instruction. These systems are growing in relevance. At their best they have been shown to achieve the same educational gain as one on one human interaction. Computer tutors have also received the attention of White House, which mentioned an tutoring platform called ASSISTments in its National Educational Technology Plan. With the fast paced adoption of these data driven systems it is important to learn how to improve the educational effectiveness of these systems by making sense of the data that is being generated from them. The studies in this proposal use data from these educational systems which primarily teach topics of Geometry and Algebra but can be applied to any domain with clearly defined sub-skills and dichotomous student response data. One of the intended impacts of this work is for these knowledge modeling contributions to facilitate the move towards computer adaptive learning in much the same way that Item Response Theory models facilitated the move towards computer adaptive testing.

Submitted by Silvana Teresinha Dornelles Studzinski (sstudzinski) on 2017-03-03T16:51:21Z No. of bitstreams: 1 Sandro José Ribeiro da Silva_.pdf: 1958508 bytes, checksum: 927c5b673859ca465e35998f946b5a64 (MD5) / Made available in DSpace on 2017-03-03T16:51:21Z (GMT). No. of bitstreams: 1 Sandro José Ribeiro da Silva_.pdf: 1958508 bytes, checksum: 927c5b673859ca465e35998f946b5a64 (MD5) Previous issue date: 2016-11-03 / Nenhuma / O desenvolvimento de habilidades de programação de sistemas computacionais é uma necessidade crescente, devido ao amplo uso de recursos computacionais nas mais diversas áreas. Ao mesmo tempo, é conhecida a deficiência existente quanto à quantidade de profissionais sendo graduados nesta área. Alguns estudos indicam dificuldades dos estudantes e ao mesmo tempo falta de metodologias adequadas como possíveis elementos contribuindo para este contexto, corroborando a necessidade de desenvolvimento de pesquisas sobre o aprendizado de linguagens de programação. Entre as possíveis soluções para este problema de motivação, o desenvolvimento de um ambiente gamificado como ferramenta de ensino para linguagens de programação vem sendo explorado em projetos de pesquisa e também em opções comerciais. Uma das deficiências observadas nestas inciativas é justamente a falta de suporte aos professores para acompanhamento da evolução dos alunos. Buscando atender esta necessidade, o presente trabalho propõe um ambiente de apoio ao ensino de lógica de programação cujo diferencial é a inclusão de recursos de análise do comportamento dos alunos, voltados para o apoio ao professor. Desta forma, o trabalho proposto alia aos jogos eletrônicos o monitoramento on-line de suas etapas, através do uso de técnicas de mineração de dados educacionais. Com base em um framework para Gamificação, foi definido e desenvolvido um ambiente Web para ensino da linguagem Lua, com aspectos de Gamificação e Mineração de Dados Educacionais. Este ambiente foi utilizado em avaliações com alunos do ensino técnico, tendo sido observados resultados promissores nos aspectos motivacionais. As avaliações envolvendo a identificação de vantagens geradas para os professores com uso dos dados sobre o comportamento dos alunos também foram positivas e indicam um bom potencial para esta abordagem. / The development of computer systems programming skills is a growing necessity, due to the wide use of computational resources in different areas. At the same time, it is known the deficiency with respect to the amount of professionals being graduated in this area. Some studies indicates difficulties of students and lack of adequate methodologies as possible elements contributing to this context, supporting the need to develop research on learning programming languages. As a possible solution to this problem of motivation, the development of a gamified environment as a teaching tool for programming languages is being explored in research projects and also commercial options. One of the deficiencies observed in these initiatives is precisely the lack of support to teachers to follow up of the evolution of students, which consists in one of the differentials of the proposed work. In this way, the work integrates to electronic games the online monitoring through the use of educational data mining techniques. Based on the framework for gamification, has been defined and developed a web environment to the Lua language teaching, with aspects of gamification and education data mining. This environment has already been tested preliminarily with technical education students, being observed promising results. A new stage of development and testing is foreseen to deepening the identification of advantages generated for teachers with the use data on the behavior of students.

This research examined teachers' online behaviors while using a digital library service--the Instructional Architect (IA)--through three consecutive studies. In the first two studies, a statistical model called latent class analysis (LCA) was applied to cluster different groups of IA teachers according to their diverse online behaviors. The third study further examined relationships between teachers' demographic characteristics and their usage patterns. Several user clusters emerged from the LCA results of Study I. These clusters were named isolated islanders, lukewarm teachers, goal-oriented brokerswindow shoppers, key brokers, beneficiaries, classroom practitioners, and dedicated sticky users. In Study II, a cleaning process was applied to the clusters discovered in Study I to further refine distinct user groups. Results revealed three clusters, key brokers, insular classroom practitioners, and ineffective islanders. In Study III, the integration of teacher demographic profiles with clustering results revealed that teaching experience and technology knowledge affected teachers' effectiveness in using the IA. The implication, contributions, and limitation of this research are discussed.

This work is about the creation of a tool for pedagogical recommendation which objective is to provide teachers, from web-based courses, personalized pedagogical recommendations generated based on the mining results of their students’ educational data. In order to guide this creation, we propose the Pedagogical Recommendation Process that counts on the coordinated work and cooperation of the Human Intelligence (domain specialists) and the Artificial Intelligence (computational tools). The process is constituted of four steps that occur in a sequential and cyclic way, starting with “Detect Practices”, where we detect if there are actions affecting the teaching and learning process. Is the next step, “Discover Patterns”, we use educational data mining techniques, based on predefined mining scenarios, to find patterns with pedagogical significance for the practices detected. In the following step, “Recommend”, it is where appropriate recommendations are offered, given the students’ current pedagogical situation. Finally, the “Monitor and Evaluate” step, where it is analyzed whether the students were positively affected by the recommendations and if they were relevant. The proposed tool was used in a case study with real data provided by a Spanish language course with 200 students enrolled, who produced more than 700 megabytes of information contained in, approximately, 1220000 triples. As results we were able to detected practices and the patterns associated to them, which were used to create recommendations, evaluated (relevance) by specialists in the educational/pedagogical domain and made available for the final users (teachers) to suggest them to their students. / A presente dissertação trata da criação de uma ferramenta para a recomendação pedagógica cujo objetivo é prover aos professores de cursos baseados na web, recomendações pedagógicas personalizadas geradas com base nos resultados da Mineração dos Dados Educacionais de seus alunos. Para orientar essa criação propomos o Processo de Recomendação Pedagógica, o qual conta com o trabalho conjunto e coordenado da Inteligência Humana (especialistas nos domínios envolvidos) e da Inteligência Artificial (ferramentas computacionais). O processo é constituído de quatro etapas que ocorrem de forma cíclica e sequencial, iniciando com “Detectar Práticas”, onde detectamos se existem ações afetando o processo de ensino e aprendizagem. Na etapa seguinte, “Descobrir Padrões”, utilizamos as técnicas de Mineração de Dados Educacionais, por meio de Cenários de Mineração predefinidos, para encontrar padrões de interesse pedagógico acerca das práticas detectadas. Na próxima etapa, “Recomendar”, são oferecidas recomendações apropriadas a atual situação pedagógica do aluno. Finalmente a etapa “Monitorar e Avaliar”, onde acompanhamos e analisamos se os alunos foram afetados positivamente pelas recomendações e se estas foram relevantes. A ferramenta de recomendação proposta foi utilizada em um estudo de caso, com dados reais provenientes de um curso de língua Espanhola com 200 alunos que produziram mais de 700 megabytes de informações dispostas em, aproximadamente, 1220000 triplas. Como resultados, fomos capazes de detectar práticas e os padrões associados a elas, que foram utilizados na criação de recomendações, avaliadas (relevância) por especialistas no domínio educacional/pedagógico, e disponibilizadas para que os usuários finais (professores) as ofereçam a seus alunos.

Submitted by Rodrigo Yoshio Tamae (rytamae@yahoo.com.br) on 2018-05-09T20:50:34Z No. of bitstreams: 1 tamae_ry_dr_prud.pdf: 8732958 bytes, checksum: adaebfe74540ed474f93e923e81fb527 (MD5) / Rejected by ALESSANDRA KUBA OSHIRO ASSUNÇÃO (alessandra@fct.unesp.br), reason: Solicitamos que realize correções na submissão seguindo as orientações abaixo: Números de página aparecem duas vezes nas folhas a partir da 140. Agradecemos a compreensão. on 2018-05-10T19:58:40Z (GMT) / Submitted by Rodrigo Yoshio Tamae (rytamae@yahoo.com.br) on 2018-05-10T20:25:22Z No. of bitstreams: 1 tamae_ry_dr_prud.pdf: 8721951 bytes, checksum: 02e3bd0d2ad16ca569a7507cc1c1583d (MD5) / Approved for entry into archive by ALESSANDRA KUBA OSHIRO ASSUNÇÃO (alessandra@fct.unesp.br) on 2018-05-11T11:39:17Z (GMT) No. of bitstreams: 1 tamae_ry_dr_prud.pdf: 8721951 bytes, checksum: 02e3bd0d2ad16ca569a7507cc1c1583d (MD5) / Made available in DSpace on 2018-05-11T11:39:18Z (GMT). No. of bitstreams: 1 tamae_ry_dr_prud.pdf: 8721951 bytes, checksum: 02e3bd0d2ad16ca569a7507cc1c1583d (MD5) Previous issue date: 2018-03-16 / Esta pesquisa de doutorado está vinculada ao Programa de Pós-Graduação em Educação da Faculdade de Ciências e Tecnologia da Universidade Estadual Paulista "Júlio de Mesquita Filho" (FCT/Unesp), campus de Presidente Prudente-SP, na linha de pesquisa "Processos Formativos, Ensino e Aprendizagem", nas áreas de Educação a distância (EaD) e Formação de Professores. O grande avanço das Tecnologias Digitais da Informação e da Comunicação (TDIC) tem provocado inúmeras mudanças em todas as áreas da ciência. Na Educação ocorre a ampla adoção e utilização dos Ambientes Virtuais de Aprendizagem (AVA), os quais podem contribuir para a utilização de TDIC, metodologias ativas de aprendizagem e que favorecem a abordagem Construcionista, Contextualizada e Significativa (CCS). A abordagem CCS é aquela em que o cursista utiliza a tecnologia como instrumento para produzir algo que parte da sua vivência e realidade, e ao se deparar com os conceitos curriculares, o professor atua como mediador para ajudá-lo a formalizar esses conceitos. Nesse contexto, a Internet e os dispositivos móveis passaram a ser utilizados em escala crescente, e tem contribuído para a proliferação de grande quantidade de dados em formato digital que, por sua vez, ainda são pouco utilizados para gerar a descoberta de conhecimento em contextos educacionais. É onde destaca-se a área de mineração de dados educacionais (MDE), que consiste no desenvolvimento de métodos e técnicas orientados a explorar tais dados digitais para melhor compreender o comportamento dos cursistas e em quais condições eles aprendem. Assim, "como utilizar técnicas de MDE para identificar indícios da abordagem CCS nos cursos da modalidade híbrida?" é a questão que norteia esta pesquisa de doutorado, pois mesmo professores qualificados para atividades docentes, muitas vezes, não possuem proficiência suficiente quanto ao uso de recursos computacionais, tais como linguagens de programação e ferramentas de banco de dados, e muito menos, quanto ao uso de técnicas de mineração de dados aplicadas à contextos educacionais. A pesquisa fez uso tanto da abordagem quantitativa quanto qualitativa, com base no delineamento metodológico Ex Post Facto ou Pesquisa não-experimental, pois o estudo foi realizado após a conclusão dos fatos. Para responder as questões norteadoras, o curso de Educação Especial na Perspectiva Inclusiva do programa Redefor/Unesp foi analisado a partir das categorias CCS (contexto do cursista, espiral de aprendizagem e ciclo de ações, aprendizagem em rede, papel do professor e formalização de conceitos) definidas com base nas indicações de Schlünzen (2000; 2015), Santos (2015) e Valente (2005). Foi utilizado o modelo de mineração de dados proposto por Fayad, Piatetsky-Shapiro e Smyth (1996) e as fases que consomem maior esforço repetitivo possibilitaram o mapeamento de padrões a serem seguidos, e para minimizar os esforços e maximizar os resultados, foi proposto e implementado um protótipo de software denominado EDMXP (Educational Data Mining eXPeriment) em linguagem de programação Java para o suporte às atividades de seleção, pré-processamento, mineração e análise de dados. As tarefas de mineração de dados utilizadas foram as de agrupamento e classificação representadas pelos algoritmos Simple KMeans, VSM e J48. Os resultados foram compilados em uma linguagem que possibilita aos profissionais de Educação melhor compreenderem os resultados (tabelas e gráficos), além de um quadro de indicadores de desempenho (dashboard). Ao final, foi possível constatar que a MDE pode ser um fator transformador em Educação a partir do momento que possibilita que se tome decisões com base em dados e em fatos, e não apenas de forma intuitiva ou por meio de experiências vivênciadas. Representa, portanto, uma nova forma de fazer e pensar a Educação. / This doctoral research is bound to the Graduate Program in Education of the Faculty of Science and Technology of the São Paulo State University "Júlio de Mesquita Filho" (FCT / Unesp), Campus of Presidente Prudente-SP, in the research line "Formative Processes, Teaching and Learning", in the areas of Distance Education (D-Learning) and Teacher Training. The great advance of the Digital Technologies of Information and Communication (DTIC) has caused fullness changes in all areas of science. In Education there is a widespread adoption and use of Virtual Learning Environment (VLE), which can contribute to the use of DTIC, active learning methodologies and favoring the Constructionist, Contextualized and Significative (CCS) approach. The CCS approach is that in which student uses technology as an instrument to produce something that arise in your own experience and reality, and when he came across with curricular concepts, teacher acts as mediator to help him to formalize these concepts. In these context, the Internet and mobile devices started to be used on a growing scale and have contributed to the proliferation of large amounts of data in digital format, which in turn are little used to generate the knowledge discovery in educational contexts. It's where stands out the area of Educational Data Mining (EDM), which consists in the development of methods and techniques designed to exploit such digital data to better understand students behavior's and in what conditions they learn. Thus, "how to use EDM techniques to identify evidence of CCS approach in hybrid mode courses?" it's the issue that guides this doctoral research, because even qualified teachers for teaching activities often lack sufficient proficiency in the use of computational resources, such as programming languages and database tools, much less regarding the use of data mining techniques applied to educational contexts. The research made use of both quantitative and qualitative approach, based on the methodological design Ex Post Facto or non-experimental research, once this study was conducted after the completion of the facts. To answer the leading questions, the Special Education course in the Inclusive Perspective of the Redefor / Unesp program was analyzed from the CCS categories (student's context, learning spiral and cycle of actions, learning network, teacher role and concepts formalization) defined according to the indications of Schlünzen (2000; 2015), Santos (2015) and Valente (2005). It was used the data mining model proposed by Fayad, Piatetsky-Shapiro and Smyth (1996) and the phases that consume most repetitive effort allowed the mapping of patterns to be followed, and to minimize efforts and to maximize results, was proposed and implemented a software prototype named EDMXP (Educational Data Mining eXPeriment) in Java programming language to support selection, preprocessing, mining and data analysis activities. The data mining tasks used were clustering and classification tasks represented by the Simple KMeans, VSM and J48 algorithms. The results were compiled in a language that enables Education professionals to better understand results (tables and graphs), as well as a dashboard of performance indicators. Finally, it was possible to verify that EDM can be a transforming factor in Education from the moment that allows decisions based on data and facts, and not only in an intuitive way or by lived experiences. It represents, therefore, a new way of doing and thinking Education.

Submitted by Marília Sousa (mariliamariafeitoza@gmail.com) on 2018-07-26T12:25:36Z No. of bitstreams: 3 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertação Marília.pdf: 1106096 bytes, checksum: 5f4d3a102f590e08a72c6af9ef02d2e4 (MD5) folha de aprovação.pdf: 114224 bytes, checksum: 83acb0aa4ff29dd5cc1364b9b391ac77 (MD5) / Approved for entry into archive by Secretaria PPGI (secretariappgi@icomp.ufam.edu.br) on 2018-07-26T18:20:47Z (GMT) No. of bitstreams: 3 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertação Marília.pdf: 1106096 bytes, checksum: 5f4d3a102f590e08a72c6af9ef02d2e4 (MD5) folha de aprovação.pdf: 114224 bytes, checksum: 83acb0aa4ff29dd5cc1364b9b391ac77 (MD5) / Approved for entry into archive by Divisão de Documentação/BC Biblioteca Central (ddbc@ufam.edu.br) on 2018-07-27T12:39:14Z (GMT) No. of bitstreams: 3 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertação Marília.pdf: 1106096 bytes, checksum: 5f4d3a102f590e08a72c6af9ef02d2e4 (MD5) folha de aprovação.pdf: 114224 bytes, checksum: 83acb0aa4ff29dd5cc1364b9b391ac77 (MD5) / Made available in DSpace on 2018-07-27T12:39:15Z (GMT). No. of bitstreams: 3 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Dissertação Marília.pdf: 1106096 bytes, checksum: 5f4d3a102f590e08a72c6af9ef02d2e4 (MD5) folha de aprovação.pdf: 114224 bytes, checksum: 83acb0aa4ff29dd5cc1364b9b391ac77 (MD5) Previous issue date: 2017-09-29 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The present work introduces the Educational Data Mining and an experiment involving prediction of partial exams. The experiment uses data of the Introduction to Computer Programming course of the Federal University of Amazonas and seeks to classify the students according to their grade, in a maximum of three classes: satisfactory, unsatisfactory and without concept (dropout students). As conclusion, there is a quantitative analysis with the predictive data. / O presente trabalho tem o intuito de apresentar a Mineração de Dados Educacionais e um experimento envolvendo previsão de provas parciais. O experimento é realizado através dos dados da disciplina de Introdução à Programação de Computadores da Universidade Federal do Amazonas e busca classificar os alunos de acordo com as notas obtidas, em no máximo três classes: satisfatório, insatisfatório e sem conceito (alunos evadidos). Como conclusão, tem-se uma análise quantitativa com os dados da previsão.