MOOC Next Week Dropout Prediction: Weekly Assessing Time and Learning Patterns

Abstract

Although Massive Open Online Course (MOOC) systems have become more prevalent in recent years, associated student attrition rates are still a major drawback. In the past decade, many researchers have sought to explore the reasons behind learner attrition or lack of interest. A growing body of literature recognises the importance of the early prediction of student attrition from MOOCs, since it can lead to timely interventions. Among them, most are concerned with identifying the best features for the entire course dropout prediction. This study focuses on innovations in predicting student dropout rates by examining their next-week-based learning activities and behaviours. The study is based on multiple MOOC platforms including 251,662 students from 7 courses with 29 runs spanning in 2013 to 2018. This study aims to build a generalised early predictive model for the weekly prediction of student completion using machine learning algorithms. In addition, this study is the first to use a 'learner's jumping behaviour' as a feature, to obtain a high dropout prediction accuracy.

Analisando a relação de conceitos de programação com o comportamento ”abuso do sistema” em programadores novatos

“…Para os classificadores simples, aplicamos os algorit-mos de árvore de decisão, como os autores utilizaram em [Baker and de Carvalho 2008], como em SVM [d Baker et al 2010], KNN e redes neurais (MLP). Em termos de algoritmos de aprendizado ensemble [Yang et al 2013], o uso de random forest tem aparecido na literatura entre as abordagens mais utilizadas para tarefas de classificac ¸ão de aprendizes em outros contextos [Alamri et al 2021]. Assim, para construir nosso modelo, empregamos vários métodos de ensemble concorrentes, como segue: Random Forest (RF) [Liu et al 2012], Gradient Boosting Machine (Gradient Boosting), [Friedman 2001] Adaptive Boosting (AdaBoost) [Hastie et al 2009] e XGBoost [Chen and Guestrin 2016] para prosseguir com a análise exploratória.…”

Predicting Student Outcomes in Online Courses Using Machine Learning Techniques: A Review

“…Several machine-learning models were examined, and LightGBM outperformed the other models. In addition, Alamri et al [75] proposed a next-week dropout-prediction model that predicts students who do not access 80% of the course content in the following week based on their learning behavior and expressed opinions in the discussion forum. Multiple classification models were tested and AdaBoost provided the highest accuracy.…”

Negotiating Learning Goals with Your Future Learning-Self

“…Machine learning models have been successfully used to model student's behavior (progress, test scores, academic performance, etc. ), based on study-related information [9,10], e.g., hours of studying, type of activities, etc. We propose utilizing a self-avatar to visualize and communicate the predictions of an AI model to the student.…”