Mining Activity Log Data to Predict Student's Outcome in a Course. Umer, R., Mathrani, A., Susnjak, T., & Lim, S. In Proceedings of the 2019 International Conference on Big Data and Education, of ICBDE'19, pages 52–58, London, United Kingdom, March, 2019. Association for Computing Machinery. ZSCC: 0000001
Mining Activity Log Data to Predict Student's Outcome in a Course [link]Paper  doi  abstract   bibtex   
Use of learning management system (LMS) is very common, which provide support to teaching staff for communication, delivery of resources and in design of learning activities. The wide spread use of technologies like LMS, provide large amount of data. Research shows that higher education institutes can make use of this data to extract data-driven insights to understand the learning process and benefit the students by supporting them in their academics. In this study we used several machine learning algorithms to predict student's outcome in a course using LMS trace data and assessment scores. Selection of the courses is based on the extent the LMS is used and is divided into two categories; distance and internal. This study confirms the importance of LMS data and assessment scores in the prediction of academic performance. However, frequent use of LMS may increase the trace data but it is not necessary improve the predictive accuracy. Predictive models developed for courses, without considering the context of use of LMS data, may not generalize the effects of LMS trace data on student's outcome in the course.
@inproceedings{umer_mining_2019,
	address = {London, United Kingdom},
	series = {{ICBDE}'19},
	title = {Mining {Activity} {Log} {Data} to {Predict} {Student}'s {Outcome} in a {Course}},
	isbn = {978-1-4503-6186-6},
	url = {https://doi.org/10.1145/3322134.3322140},
	doi = {10/ggx75h},
	abstract = {Use of learning management system (LMS) is very common, which provide support to teaching staff for communication, delivery of resources and in design of learning activities. The wide spread use of technologies like LMS, provide large amount of data. Research shows that higher education institutes can make use of this data to extract data-driven insights to understand the learning process and benefit the students by supporting them in their academics. In this study we used several machine learning algorithms to predict student's outcome in a course using LMS trace data and assessment scores. Selection of the courses is based on the extent the LMS is used and is divided into two categories; distance and internal. This study confirms the importance of LMS data and assessment scores in the prediction of academic performance. However, frequent use of LMS may increase the trace data but it is not necessary improve the predictive accuracy. Predictive models developed for courses, without considering the context of use of LMS data, may not generalize the effects of LMS trace data on student's outcome in the course.},
	urldate = {2020-06-03},
	booktitle = {Proceedings of the 2019 {International} {Conference} on {Big} {Data} and {Education}},
	publisher = {Association for Computing Machinery},
	author = {Umer, Rahila and Mathrani, Anuradha and Susnjak, Teo and Lim, Suriadi},
	month = mar,
	year = {2019},
	note = {ZSCC: 0000001},
	keywords = {Classification, Education data mining, Learning analytics, prediction},
	pages = {52--58}
}
Downloads: 0