Learner2Vec-Based Learner Community Evolution Analysis- A Case Study Involving Student Card Data

Abstract

The learner community evolution analysis is an important technique in the area of social network analysis and education management. Accurately expressing the complex characteristics of individual social behavior is the core problem associated with analyzing learner community evolution. In this paper, we propose an approach to learner community evolution analysis. First, we present a vector representation model of learner behavior. Then, we use the vector distance calculation method to construct a relational network of learners. Third, the time slice division and the Louvain community discovery algorithm are used to analyze the evolution of the learner community. The experimental results demonstrate that the proposed approach can effectively describe the vector features of learners' individuality and accurately detect factors influencing learner community evolution, such as class grouping, postgraduate entrance examination preparation, and job hunting.