A Group Recommender System for Students

Abstract

In a project-based course, students must spend a whole semester to finish a group project. If a student is not familiar with other classmates in the course, it is difficult for the student to form a group. To this end, this project aims to develop a recommender for students to help them form groups. Given historical student records with their demographic data, academic data, and course performance data, features, patterns and relations are extracted from the data and machine learning techniques are used to build a model and recommend group members for a student with the objective of achieving an expected grade for the group project. This project has several long-term impacts: helping students form groups based on their expected results, improving students' learning experiences, and promoting the application of machine learning techniques on education. The group recommender system will be integrated in the online project-based course management system (PCMS). PCMS is being developed by the PI sponsored by another TDG grant.