abstract: Programming is quickly becoming as ubiquitous and essential a skill as general mathematics. However, many elementary and high school students are still not aware of what the computer science field entails. To make matters worse, students who are introduced to computer science are frequently being fed only part of what it is about rather than its entire construction. Consequently, they feel out of their depth when they approach college. Research has discovered that by teaching computer science and programming through a problem-driven approach and focusing on a combination of syntax and computational thinking, students can be prepared when entering higher levels of computer science education. This thesis describes the design, development, and early user testing of a theory-based virtual world for computer science instruction called System Dot. System Dot was designed to visually manifest programming instructions into interactable objects, giving players a way to see coding as tangible entities rather than text on a white screen. In order for System Dot to convey the true nature of computer science, a custom predictive recursive descent parser was embedded in the program to validate any user-generated solutions to pre-defined logical platforming puzzles. Steps were taken to adapt the virtual world to player behavior by creating a system to detect their learning style playing the game. Through a dynamic Bayesian network, System Dot aims to classify a player’s learning style based on the Felder-Sylverman Learning Style Model (FSLSM). Testers played through the first half of System Dot, which was enough to test out the Bayesian network and initial learning style classification. This classification was then compared to the assessment by Felder’s Index of Learning Styles Questionnaire (ILSQ). Lastly, this thesis will also discuss ways to use the results from the user testing to implement a personalized feedback system for the virtual world in the future and what has been learned through the learning style method. / Dissertation/Thesis / Masters Thesis Computer Science 2017

Com efeito, diversos estudos nas áreas de psicologia cognitiva e pedagogia apontam que cada indivíduo possui diferentes maneiras de captar, processar, analisar e organizar informações durante o processo de aprendizado, o que fundamenta o conceito de Estilos de Aprendizagem (EA). Em vista disso, diversos sistemas educacionais adaptativos foram propostos com o intuito de proporcionar conteúdo personalizado em seus cursos. Porém, em boa parte dos casos, estes sistemas fazem uso de questionários para identificar os estilos de aprendizagem, e este método pode mostrar-se inviável em algumas situações. Isso ocorre pois o preenchimento dos questionários demanda um esforço adicional por parte do aluno, além fazer uma avaliação estática dos EA, desconsiderando possíveis variações em suas preferências ao longo do tempo. Supõe-se que uma estratégia de detecção automática e dinâmica dos EA baseada no comportamento dos estudantes pode ser mais proveitosa neste sentido, pois é isenta destas limitações. Deste modo, a proposta neste trabalho é investigar diferentes técnicas relacionadas ao aprendizado de máquina (especialmente algoritmos de classificação) aplicadas à predição automática dos estilos de aprendizagem de estudantes, a partir de suas interações em um ambiente virtual de ensino. Dentre os inúmeros modelos de EA propostos na literatura, optou-se por usar o modelo de Felder-Silverman como base para os experimentos. Como estudo de caso, foram rastreadas as interações de 105 estudantes de um curso de pós-graduação em fonoaudiologia ministrado integralmente pelo sistema Moodle. Além disso, estes alunos foram solicitados a responder ao questionário ILS, o qual indica a preferência de cada indivíduo de acordo com o modelo de Felder-Silverman. Para a construção dos conjuntos de dados, foram coletadas informações como a quantidade de visitas, tempo gasto e interação dos usuários em cada tipo de recurso (recursos de vídeo, formulários de avaliação, fórum, etc.). Estes conjuntos de dados no formato atributo-valor serviram de entrada para quatro algoritmos de classificação: Naïve Bayes, aprendizado baseado em instâncias (kNN), Redes Neurais Artificiais (MultiLayer Perceptron) e Árvores de Decisão (J48), combinados com métodos de seleção de atributos e executados em validação cruzada. Para fins de experimentação, foram avaliadas as taxas de acertos e erros dos algoritmos em relação aos resultados apontados pelo questionário ILS, em cada umas das dimensões do modelo de Felder-Silverman. Os resultados apontaram para o uso de mais de um classificador - Naïve Bayes e aprendizagem baseada em instância - dependendo da dimensão do estilo de aprendizagem. A metodologia aplicada foi comparada com sete trabalhos correlatos da literatura; Os resultados demonstraram uma performance superior aos trabalhos anteriores em quase todas as dimensões. Portanto, o presente trabalho contribui para o contexto da informática aplicada à educação, especificamente no que diz respeito à implementação de sistemas educacionais adaptativos, com base em uma análise comparativa entre diferentes técnicas aplicadas ao mesmo problema. Sendo assim, as conclusões obtidas devem colaborar para o aprimoramento do processo de modelagem de estudantes. Além disso, são levantadas discussões a respeito dos resultados, que podem auxiliar na direção de futuros trabalhos da área. / In fact, several studies in the areas of cognitive psychology and pedagogy point out that each individual has different ways of capturing, processing, analyzing and organizing information during the learning process, which supports the concept of Learning Styles (LS). Therefore, several adaptive educational systems were proposed with the aim of providing personalized content in their courses. However, in most cases, these systems use questionnaires to identify learning styles, and this method may prove unfeasible in some situations. This is because filling in the questionnaires requires an additional effort on the part of the student, besides, this approach makes a static evaluation of the LS, disregarding possible variations in their preferences over time. It is assumed that an automatic and dynamic detection of LS based on student behavior may be more useful in this sense, since it is exempt from these limitations. In this way, the proposal in this work is to investigate different techniques related to machine learning (especially classification algorithms) applied to the automatic prediction of student learning styles, based on their interactions in a virtual teaching environment. Among the many LS models proposed in the literature, we chose to use the Felder-Silverman model (FSLSM). As a case study, the interactions of 105 students from a post-graduate course in speech therapy were studied. In addition, these students were asked to respond to the ILS questionnaire, which indicates the preference of each individual according to FSLSM. In order to construct the data sets, information was collected such as the number of visits, time spent and user interaction in each type of resource (video resources, evaluation forms, forum, etc.). These data sets in the attribute-value format served as input to four classification algorithms: Naïve Bayes, instance-based learning (kNN), MultiLayer Perceptron and Decision Trees (J48), combined with attribute selection methods and executed in cross-validation. For the experimentation, the accuracy and error rates of the algorithms were evaluated in relation to the results indicated by the ILS questionnaire, in each one of FSLSM dimensions. Our results pointed out to the use of more than one classifier, Naïve Bayes and Instance-based Learning, depending on the learning style dimension. We compared our methodology to seven works of the literature; the results demonstrated a performance superior to the previous works in almost every dimension. The present work contributes to the context of informatics applied to education, specifically with regard to the implementation of adaptive educational systems, based on a comparative analysis of different methods applied to the same problem. Therefore, the conclusions obtained should contribute to the improvement of the student modeling process. In addition, discussions are held regarding the results, which may assist in the direction of future work in this area.

Com efeito, diversos estudos nas áreas de psicologia cognitiva e pedagogia apontam que cada indivíduo possui diferentes maneiras de captar, processar, analisar e organizar informações durante o processo de aprendizado, o que fundamenta o conceito de Estilos de Aprendizagem (EA). Em vista disso, diversos sistemas educacionais adaptativos foram propostos com o intuito de proporcionar conteúdo personalizado em seus cursos. Porém, em boa parte dos casos, estes sistemas fazem uso de questionários para identificar os estilos de aprendizagem, e este método pode mostrar-se inviável em algumas situações. Isso ocorre pois o preenchimento dos questionários demanda um esforço adicional por parte do aluno, além fazer uma avaliação estática dos EA, desconsiderando possíveis variações em suas preferências ao longo do tempo. Supõe-se que uma estratégia de detecção automática e dinâmica dos EA baseada no comportamento dos estudantes pode ser mais proveitosa neste sentido, pois é isenta destas limitações. Deste modo, a proposta neste trabalho é investigar diferentes técnicas relacionadas ao aprendizado de máquina (especialmente algoritmos de classificação) aplicadas à predição automática dos estilos de aprendizagem de estudantes, a partir de suas interações em um ambiente virtual de ensino. Dentre os inúmeros modelos de EA propostos na literatura, optou-se por usar o modelo de Felder-Silverman como base para os experimentos. Como estudo de caso, foram rastreadas as interações de 105 estudantes de um curso de pós-graduação em fonoaudiologia ministrado integralmente pelo sistema Moodle. Além disso, estes alunos foram solicitados a responder ao questionário ILS, o qual indica a preferência de cada indivíduo de acordo com o modelo de Felder-Silverman. Para a construção dos conjuntos de dados, foram coletadas informações como a quantidade de visitas, tempo gasto e interação dos usuários em cada tipo de recurso (recursos de vídeo, formulários de avaliação, fórum, etc.). Estes conjuntos de dados no formato atributo-valor serviram de entrada para quatro algoritmos de classificação: Naïve Bayes, aprendizado baseado em instâncias (kNN), Redes Neurais Artificiais (MultiLayer Perceptron) e Árvores de Decisão (J48), combinados com métodos de seleção de atributos e executados em validação cruzada. Para fins de experimentação, foram avaliadas as taxas de acertos e erros dos algoritmos em relação aos resultados apontados pelo questionário ILS, em cada umas das dimensões do modelo de Felder-Silverman. Os resultados apontaram para o uso de mais de um classificador - Naïve Bayes e aprendizagem baseada em instância - dependendo da dimensão do estilo de aprendizagem. A metodologia aplicada foi comparada com sete trabalhos correlatos da literatura; Os resultados demonstraram uma performance superior aos trabalhos anteriores em quase todas as dimensões. Portanto, o presente trabalho contribui para o contexto da informática aplicada à educação, especificamente no que diz respeito à implementação de sistemas educacionais adaptativos, com base em uma análise comparativa entre diferentes técnicas aplicadas ao mesmo problema. Sendo assim, as conclusões obtidas devem colaborar para o aprimoramento do processo de modelagem de estudantes. Além disso, são levantadas discussões a respeito dos resultados, que podem auxiliar na direção de futuros trabalhos da área. / In fact, several studies in the areas of cognitive psychology and pedagogy point out that each individual has different ways of capturing, processing, analyzing and organizing information during the learning process, which supports the concept of Learning Styles (LS). Therefore, several adaptive educational systems were proposed with the aim of providing personalized content in their courses. However, in most cases, these systems use questionnaires to identify learning styles, and this method may prove unfeasible in some situations. This is because filling in the questionnaires requires an additional effort on the part of the student, besides, this approach makes a static evaluation of the LS, disregarding possible variations in their preferences over time. It is assumed that an automatic and dynamic detection of LS based on student behavior may be more useful in this sense, since it is exempt from these limitations. In this way, the proposal in this work is to investigate different techniques related to machine learning (especially classification algorithms) applied to the automatic prediction of student learning styles, based on their interactions in a virtual teaching environment. Among the many LS models proposed in the literature, we chose to use the Felder-Silverman model (FSLSM). As a case study, the interactions of 105 students from a post-graduate course in speech therapy were studied. In addition, these students were asked to respond to the ILS questionnaire, which indicates the preference of each individual according to FSLSM. In order to construct the data sets, information was collected such as the number of visits, time spent and user interaction in each type of resource (video resources, evaluation forms, forum, etc.). These data sets in the attribute-value format served as input to four classification algorithms: Naïve Bayes, instance-based learning (kNN), MultiLayer Perceptron and Decision Trees (J48), combined with attribute selection methods and executed in cross-validation. For the experimentation, the accuracy and error rates of the algorithms were evaluated in relation to the results indicated by the ILS questionnaire, in each one of FSLSM dimensions. Our results pointed out to the use of more than one classifier, Naïve Bayes and Instance-based Learning, depending on the learning style dimension. We compared our methodology to seven works of the literature; the results demonstrated a performance superior to the previous works in almost every dimension. The present work contributes to the context of informatics applied to education, specifically with regard to the implementation of adaptive educational systems, based on a comparative analysis of different methods applied to the same problem. Therefore, the conclusions obtained should contribute to the improvement of the student modeling process. In addition, discussions are held regarding the results, which may assist in the direction of future work in this area.

This study investigates existing learning management systems practices, in this case Canvas and Moodle in relation to user personalization and students’ learning styles as both factors are closely contribute to the design of a meaningful learning experience for learners. With the expansion of these teaching tools and methods, it seems crucial to determine to what extent they actually serve the learner and what role is really given to the student using these online platforms. Factors such as instructors’ feedback, peer communication, learning objects and follow-up will be examined. This study is anchored in a Swedish academic setting, and aims to provide a comprehensive overview of learners' needs, expectations, and preferences to benefit educational institutions as well as LMS developers. The goal is to assess how such factors play an essential role in the personalization of learning tools and to suggest that their consideration can lead to the development of more intuitive LMS platforms that do not solely rely on content uploaded by teachers, but that can in turn potentially offer relevant content tailored to each user. / Den här uppsatsen undersöker befintliga praxis för lärande hanteringssystem, i detta fall Canvas och Moodle, i förhållande till användaranpassning och studenternas inlärningsstilar, eftersom båda faktorerna bidrar till utformningen av en meningsfull inlärningsupplevelse för studenterna. På grund av expansionen av dessa undervisningsverktyg verkar det avgörande att bestämma i vilken utsträckning de faktiskt tjänar inläraren och vilken roll studenten verkligen får när hen använder dessa plattformar. Faktorer som lärarnas återkoppling, kommunikation med andra elever, lärandeobjekt och uppföljning kommer att undersökas noggrant. Studien är förankrad i en svensk akademisk miljö och syftar att ge en heltäckande översikt av inlärarnas behov, förväntningar och preferenser. Målet är att förstå hur dessa faktorer spelar en väsentlig roll i personaliseringen av lärverktyg och att föreslå att deras beaktande kan leda till utveckling av mer intuitiva LMS-plattformar som inte enbart förlitar sig på innehåll som laddas upp av lärare, utan som i sin tur potentiellt kan erbjuda relevant innehåll som är skräddarsytt för varje användare.

Αντικείμενο της συγκεκριμένης εργασίας αποτελεί η μελέτη των χαρακτηριστικών του μοντέλου εκπαιδευόμενου, ο καθορισμός των απαιτούμενων εξ' αυτών για ένα προσαρμοστικό σύστημα ηλεκτρονικής μάθησης καθώς και ο αποδοτικότερος τρόπος αναπαράστασής τους. Επίσης μελετώνται οι τρόποι συλλογής δεδομένων για την δημιουργία και συνεχή ανανέωση του μοντέλου (μοντελοποίηση) καθώς και οι αλγόριθμοι που χρησιμοποιούνται, με σκοπό την υλοποίηση του καταλληλότερου και αποδοτικότερου εξ' αυτών. Η διπλωματική στοχεύει στη δημιουργία ενός μοντέλου χρήστη-εκπαιδευόμενου συνοδευόμενο από τους αντίστοιχους αλγορίθμους μοντελοποίησης το οποίο θα συγκεντρώνει την απαραίτητη πληροφορία για την αυτοματοποιημένη και προσαρμοσμένη επιλογή κατάλληλου εκπαιδευτικού υλικού προς τον συγκεκριμένο χρήστη, με ευκολία και αφαιρετικότητα προς τον χρήστη. / In the present thesis we are studying the characteristics that can be represented in a user model, we define those that are essential in a student model and we propose an efficient way of representation. We also study the different ways of constructing and updating a student model and the algorithms that can be used in order to implement the most appropriate and efficient of them. The main purpose of the thesis is the development of a student model, with the suitable algorithms, that will support the automated process of adapted educational material provision in an easy and abstract way.

Philosophiae Doctor - PhD / This study attempts to understand how distance education practices influence the learning of calculus. Understanding student learning in a distance education environment is an important factor to consider in improving the learning experiences of those students who for one reason or the other opt not to study in conventional institutions of higher education. On one hand, understanding student learning may illuminate the influences that the learning environment has on student learning and on the other hand, it may inform on how learning experiences can be improved. The aim of this study is to acquire a deeper understanding of the diverse manner in which distance students learn calculus. Specific focus is also placed on how the distance education context of the Zimbabwe Open University (ZOU) influences student learning. The study describes a group of students' experiences of learning calculus in the ZOU distance education environment. The study also describes the students' learning styles and relates these to their mathematical understanding of certain calculus concepts. The specific content topics of "limit of function" and "derivative of function" are used to view achievement and performance, thereby indicating the distance students' mathematical understanding. The information processing learning theory is used as the theoretical framework for this study. The constructs of learning styles and mathematical understanding are used to illuminate the student's learning processes. The study used the Felder-Silverman learning styles model and Hiebert and Carpenter's notion of mathematical understanding to expound these constructs. The distance education environment of the B.Sc. Mathematics and Statistics (BSMS) programme at the ZOU provided the context of the study and an interpretive case study approach was adopted. A group of students registered in a first year first semester calculus course were studied. Data were collected from students based in four ZOU regional centres; namely Harare, Mashonaland Central, Mashonaland West, and Masvingo. These regional centres were conveniently selected for the study on the basis of proximity and accessibility. A total sample of twenty six students was involved and data for the in-depth part of the study emanated from five students who were purposively selected to participate in interviews. The interviewees were selected on the basis of their performance in a written calculus test. Data for this study were collected through use of learning journals, learning styles preference questionnaires, calculus tests and interviews. The data on students' learning experiences were predominantly qualitative in nature though supported by some quantitative data. The data on learning styles and mathematical understanding were also qualitatively analysed and presented case by case for the five interviewees. The study established that in a distance education system, the type of learning environment has the potential to influence students' learning, both positively and negatively, of which the main contributing factor is the learning support system. The study found that the learning support system provided by the institution and distance educators can have an impact on student learning. With reference to the calculus course in the BSMS programme, the study identified specific aspects where the environment facilitated or deterred learning. The study also revealed that students have varied learning style preferences, and that the learning environment has the potential to impact on students' learning styles. Since learning styles occupy a central place when it comes to improving distance learning materials, the study further explored the relationship between the constructs of learning styles and mathematical understanding. The study revealed that students' learning styles can influence the students' mathematical understanding. Improving students' learning in a distance education environment rests mainly on improving the learning materials and the support systems. A carefully designed and well supported instructional distance learning package can facilitate learning. Implications of the findings point towards the improvement of the distance teaching processes through the improvement of learning materials and the learning support systems for the BSMS distance education programme.

Submitted by Jos? Henrique Henrique (jose.neves@ufvjm.edu.br) on 2017-01-09T12:21:59Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) andre_vinicius_gon?alves.pdf: 1266538 bytes, checksum: 42c3fe90b9d66c8cb7b901a10e548f1b (MD5) / Approved for entry into archive by Rodrigo Martins Cruz (rodrigo.cruz@ufvjm.edu.br) on 2017-01-31T13:56:36Z (GMT) No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) andre_vinicius_gon?alves.pdf: 1266538 bytes, checksum: 42c3fe90b9d66c8cb7b901a10e548f1b (MD5) / Made available in DSpace on 2017-01-31T13:56:36Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) andre_vinicius_gon?alves.pdf: 1266538 bytes, checksum: 42c3fe90b9d66c8cb7b901a10e548f1b (MD5) Previous issue date: 2016-06 / Nos ?ltimos dez anos muitos pesquisadores t?m realizado estudos sobre assist?ncia personalizada e inteligente em Ambientes Educacionais a Dist?ncia, baseada na identifica??o dos Estilos de Aprendizagem. Sabe-se que o aprendizado ? algo extremamente particular, pois cada estudante possui estilos pr?prios e pode sofrer mudan?as diante de situa??es diversas como, por exemplo, objetivo, motiva??o, personalidade, etc. Por isso, o conceito de adaptabilidade do conte?do did?tico tem se tornado de grande import?ncia na personaliza??o do Sistema de Gerenciamento de Aprendizagem (SGA). Diante desse fato, Dor?a (2012) prop?e uma abordagem de Sistema Adaptativo e Inteligente para Educa??o (SAIE), utilizando t?cnicas probabil?sticas e Intelig?ncia Artificial (IA), capaz de detectar e adaptar, de maneira din?mica e autom?tica, os estilos de aprendizagem do estudante, considerando o Modelo de Estilo de Aprendizagem Felder-Silverman?s. Ap?s pesquisa detalhada, foram propostas algumas adapta??es baseadas na abordagem original, alterando o funcionamento de dois componentes espec?ficos: o M?dulo Pedag?gico e o Componente de Modelagem do Estudante. Al?m disso, prop?e-se uma nova estrutura do Modelo Estudante, contemplando o hist?rico de desempenho do aluno nos processos avaliativos. Por conseguinte, realizaram-se testes para avaliar os impactos de tais mudan?as por meio uma compara??o estat?stica utilizando o m?todo T-Pareado. Pelos resultados obtidos, as ideias deste trabalho proporcionaram uma melhora m?dia de 6,07% no desempenho avaliativo do estudante e uma redu??o m?dia de 68,27% nos problemas de aprendizagem, demonstrando efici?ncia e efic?cia da proposta. / Disserta??o (Mestrado Profissional) ? Programa de P?s-Gradua??o em Educa??o, Universidade Federal dos Vales do Jequitinhonha e Mucuri, 2015. / Since last decade many researchers have been conducting studies on personalized and intelligent assistance in distance education based on identification of learning styles. It is known that learning is something very particular because each student has their own styles and are subject to change on a variety of situations such as goal, motivation, personality, etc. Therefore, this study discusses the concept of adaptability of educational content as a way to provide customization of Learning Management System (LMS). Through probabilistic techniques and Artificial Intelligence (AI), Dor?a (2012) proposed a approach Adaptive and Intelligent System for Education (AIES) able to dynamically and automatically detect, select and adapt learning objects based on the student?s profile through Felder-Silverman Learning Styles Model (FSLSM). After detailed study, it has been proposed some adaptations based on this approach, thereby altering the operation of two specific components: the Pedagogical Module and the Student Modeling Component. In addition, it is proposed a new structure Model Student, considering learner performance history in the evaluation processes. Therefore, it carried out tests to assess the impacts of such changes through a statistical comparison by T-Paired method. From the results, the ideas in this work provides an average improvement of 6.07% in the performance evaluation of the student and an average reduction of 68.27% in the learning problems, demonstrating proposal of efficiency and effectiveness.