Consideration of the local correlation of learning behaviors to predict dropouts from MOOCs

Abstract

Recently, Massive Open Online Courses (MOOCs) have become a major online learning methodology for millions of people worldwide. However, the dropout rates from several current MOOCs are high. Usually, dropout prediction aims to predict whether a learner will exhibit learning behaviors during several consecutive days in the future. Therefore, the information related to the learning behaviors of a learner in several consecutive days should be considered. After in-depth analysis of the learning behavior patterns of the MOOC learners, this study reports that learners often exhibit similar learning behaviors on several consecutive days, i.e., the learning status of a learner for the subsequent day is likely to be similar to that for the previous day. Based on this characteristic of MOOC learning, this study proposes a new simple feature matrix for keeping information related to the local correlation of learning behaviors and a new Convolutional Neural Network (CNN) model for predicting the dropout. Extensive experimental validations illustrate that the local correlation of learning behaviors should not be neglected. The proposed CNN model considers this characteristic and improves the dropout prediction accuracy. Furthermore, the proposed model can be used to predict dropout temporally and early when sufficient data are collected.

Determining Factors in MOOCs Completion Rates: Application Test in Energy Sustainability Courses

“…In response to the general objective of this research, it was found that personal factors [3,4], family factors [8][9][10], social factors [14], instructional design factors [17,18] and labor factors [29] influence both the decision to take a MOOC (expectation-value) and, to a certain extent, engagement in the training program-measured by completion rates-as noted in the review of the state of the art. However, not all factors do so in the same way.…”

“…Instructional design factors are critical to the commitment and completion rates. Skills such as motivation for achievement and self-esteem, self-efficacy, effectiveness, design, and development are crucial elements to consider in the instructional design [17] and, since they are new learning models, studies should be conducted to learn in detail about the changes that students make [18]. Research has learned about the effects of instructional designs that can be useful for MOOCs' providers and students in their efforts to develop strategies to increase completion rates of MOOCs [19], as well as to determine the duration and analyze group behaviors [20] in order to implement different ways to increase the percentage of effectiveness.…”

MOOC Dropout Prediction Using FWTS-CNN Model Based on Fused Feature Weighting and Time Series

“…Considering the good performance of deep learning, some researchers have used time series data as input data for deep learning to improve the accuracy of dropout prediction. For example, Wen et al [37] designed a two-dimensional matrix based on time series as an input to the CNN model, combining time information with the learner's behavioral features to solve the problem of local correlation of behavioral features using time series. Qiu et al [15] proposed an end-to-end dropout prediction model based on convolutional neural networks that integrates feature extraction and classification into a single framework.…”

Processo de Mineração de Dados Educacionais aplicado na Previsão do Desempenho de Alunos: Uma comparação entre as Técnicas de Aprendizagem de Máquina e Aprendizagem Profunda

“…Guo et al (2019) chegaram a resultados que superam a precisão de soluções presentes no estado da arte em até 2,4%, e concluíram que sua pesquisa pode auxiliar professores e tutores a priorizar suas respostas e gerenciar melhor várias postagens, de modo que esses profissionais da educação possam responder às perguntas dos alunos em tempo hábil e ajudar a reduzir as taxas de evasão nesses cursos. Wen et al (2020) também utilizaram um modelo de AP para pesquisa no âmbito educacional, com o objetivo de identificar antecipadamente a desistência em Massive Open Online Courses (MOOCs). Nesse sentido, depois que Wen et al (2020) realizaram uma análise dos padrões de comportamento de aprendizagem dos alunos de um MOOC, relataram que esses estudantes geralmente exibem comportamentos de aprendizagem semelhantes em vários dias consecutivos (o status de aprendizagem de um aluno para o dia subsequente, provavelmente será semelhante ao do dia anterior).…”

“…Wen et al (2020) também utilizaram um modelo de AP para pesquisa no âmbito educacional, com o objetivo de identificar antecipadamente a desistência em Massive Open Online Courses (MOOCs). Nesse sentido, depois que Wen et al (2020) realizaram uma análise dos padrões de comportamento de aprendizagem dos alunos de um MOOC, relataram que esses estudantes geralmente exibem comportamentos de aprendizagem semelhantes em vários dias consecutivos (o status de aprendizagem de um aluno para o dia subsequente, provavelmente será semelhante ao do dia anterior). Embasados nessa premissa Wen et al (2020) propuseram uma base de dados formada por atributos relacionados à correlação local de comportamentos de Souza, V. F., Santos, T. C. B. RBIE v.29 -2021 529 aprendizagem, sobre a qual aplicaram uma Rede Neural Convolucional, gerando um modelo para prever o abandono de alunos em MOOCs.…”