Data mining techniques to improve predictions accuracy of students' academic performance : a case study with Xorro-Q : a thesis presented in partial fulfilment of the requirements for Master of Information Science (IT) at Massey University, Auckland, New Zealand in 2018
Loading...
Date
2018
DOI
Open Access Location
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Massey University
Rights
The Author
Abstract
Recent research in analytics has assisted policy makers capitalize on their ever-increasing data repositories and make data-driven predictions to create a vision for developing strategies to achieve their business targets. This is especially relevant in educational environments where data mining techniques can be applied to make predictions around students’ academic performance. This can help educators align a teaching strategy which encourages and assists students with their learning. Suit-able pedagogical support can be provided to enhance the overall student learning experience.
This study is in the educational domain where student-related course data has been used to extract insights on student performances over the study period. Exten-sive data collected from an educational tool (Xorro-Q) used in an engineering course delivery has aided this investigation. Data collected from Xorro-Q comprised stu-dent scores from real-time and self-paced activities set by educators over a 12-week semester period along with students’ final Exam scores and scores from a compulsory prerequisite course. Popular data mining techniques have been applied to predict the academic performance of students based on data extracted from Xorro-Q. This is done by training the classifier using four different algorithms, namely, Naive Bayes, Logistic regression, K nearest neighbour and Random Forest. Process mining techniques have been applied along with the general features to find out the effectiveness, such as improvement in accuracy of predictions. The study has further implications in enhancing value of the role of analytics for predictive modelling by incorporating process mining features in the training set of data.
Description
Figures 2.1 (p.8), 2.3 (p.17), 3.1 (p.26), 3.2 (p.28), 3.3 (p.30), 3.5 (p.31), 3.6 (p.33), 3.9 (p.36) & 3.10 (p.37) have been removed for copyright reasons. They may be accessed using the sources listed in the bibliography.
Keywords
Data mining -- Computer programs, Academic achievement -- Data processing, Prediction of scholastic success