Predicting the risk of attrition for undergraduate students with time based modelling
|dc.identifier.citation||Chai, K. and Gibson, D. 2015. Predicting the risk of attrition for undergraduate students with time based modelling, pp. 109-116.|
Improving student retention is an important and challenging problem for universities. This paper reports on the development of a student attrition model for predicting which first year students are most at-risk of leaving at various points in time during their first semester of study. The objective of developing such a model is to assist universities by proactively supporting and retaining these students as their situations and risk change over time. The study evaluated different models for predicting student attrition at four different time periods throughout a semester study period: pre-enrolment, enrolment, in-semester and end-of-semester models. A dataset of 23,291 students who enrolled in their first semester between 2011-2013 was extracted from various data sources. Three supervised machine learning techniques were tested to develop the predictive models: logistic regression, decision trees and random forests. The performance of these models were evaluated using the precision and recall metrics. The model achieved the best performance and user utility using logistic regression (67% precision, 29% recall). A web application was developed for users to visualise and interact with the model results to assist in the targeting of student intervention responses and programs.
|dc.title||Predicting the risk of attrition for undergraduate students with time based modelling|
|dcterms.source.title||Proceedings of the 12th International Conference on Cognition and Exploratory Learning in the Digital Age, CELDA 2015|
|dcterms.source.series||Proceedings of the 12th International Conference on Cognition and Exploratory Learning in the Digital Age, CELDA 2015|
|curtin.department||Curtin Teaching and Learning (CTL)|
|curtin.accessStatus||Fulltext not available|
Files in this item
There are no files associated with this item.