Ftgws: Forming Optimal Tutor Group For Weak Students Discovered In Educational Settings

The task of experts discovering, as one of the most important research issues in social networks, has been widely studied by many researchers in recent years. However, there are extremely few works considering this issue in educational settings. In this work, we focus on the problem of forming tutor group for weak students based on their knowledge state. To solve this problem, a novel framework based on Student-Skill Interaction (SSI) model and set covering theory is proposed, which is called FTGWS. The FTGWS framework contains three major steps: firstly, building SSI models for each student and each skill he or she has encountered; then, discovering the top-k weak students based on their knowledge state; finally, forming the optimal tutor group for each weak student. We evaluate our framework on a real-word dataset which contains 28834 students and 244 skills. The experiments show that the framework is capable of producing high-quality solutions (for 93% of weak students, the size of the optimal tutor group can be decreased up to 2 students).