Submit a Proposal for a Presentation

Note: The role you are currently using is a different role from your usual role. The messages below are those that would appear if you were actually logged in as a user with that role. The messages are displayed for testing purposes, even though they may not strictly be relevant to your actual login status.

In order to add, edit or delete a presentation proposal for this conference, you need to log in to this website and enrol in this Moodle course.

If you have a username but are not enrolled in this Moodle course, please click here to enrol and then revisit this page to continue the proposal submission process.

To add a presentation proposal, please click the "Add entry" link below, and fill out the form on the next page.

To edit a presentation proposal, please click the "View single" link below, and then click the edit icon on the next page.

To delete a presentation proposal, please click the "View single" link below, and then click the delete icon on the next page.

ID : 1837
Original submission元の投稿 :
Title題名 : Predicting Student Performance Using Moodle Activity Log Data
Type種類 : Presentation (20 mins)
Category部類 : Individual presentation
Language言語 : English
Abstract要約 :

"Digital Learning and Massive Data Analysis" is one of the topics of current student learning assessment in various universities. In Ming Chuan University, our Moodle platform has been used since 2006 and has accumulated more than 46 million logins. It has more than 60,000 courses, and hundreds of millions of student activity logs. These precious and huge amounts of data can be mined through data exploration technologies. In this paper, we used each student's activity logs in Moodle and the student's final semester grades to divide the 18-weeks student learning activity data into six study periods. Models were established every 3 weeks, 6 weeks, 9 weeks, 12 weeks, and 18 weeks. In this study,  classification algorithms used with feature selection are Naïve Bayes, MultilayerPerceptron, Logistic, J48 and RandomForest. The classification model of each interval was used to predict whether students in the new semester would pass or fail the course. The accuracy of the predictions of this study is about 75%. In this forecasting result, it is mainly expected to let teachers understand the learning status of each student at each stage, so as to provide early warning and assistance.

Keywordsキーワード : Moodle, learning assessment, classification, educational data mining
Topicsトピック :
Presentation times発表時間 :
Main presenter筆頭発表者 : Wen-Chu KUO
Affiliation所属 : Ming Chuan University (Taipei, Taiwan)
EmailEメール :
Co-presenter (2)共同研究者(2) : Ms. Wen-Yu CHEN (Ming Chuan University)
Handout資料 :
Slidesスライド : PDF document MoodleJPY_2019.02.25.pdf
Commentsコメント :
Added by入力ユーザ :
Date added入力日付 : Wednesday, 14 November 2018, 12:36 PM
Date modified更新日付 : Monday, 25 February 2019, 12:12 PM
Message from the
Vetting Committee
Peer review査読 :
Submission status処理状況 :
Schedule numberスケジュール番号 : 308-P
Schedule day開催予定日 : Mar 1st (Fri)
Schedule time開催予定時刻 : 10: 10 - 10: 20
Schedule duration開催予定時間 : 10 mins
Schedule room name開催予定室名 : Room 202 (80 seats)
Schedule room seats開催予定室名の坐整数 : 80 seats
Schedule audience聴講予定者 :