A Recommender Model Using Social Tie Strength for the Chunk Learning System
Abstract
With the onset of COVID-19, rising tuition costs, and technological advancements, online courses have become a pervasive medium through which education is conducted. Currently, several online educational services tailor education to students through various methods of recommender models. One such system, the Curated Heuristic Using a Network of Knowledge (CHUNK) Learning, developed at the Naval Postgraduate School, uses a recommender system that relies on user profile attributes. We propose a complementary recommendation system to expand upon CHUNK's current recommender method by incorporating implicit recommendations from a user's social network based on tie strength between learners. In this work, we create a synthetic social network of learners and calculate the Jaccard Index and Pearson Correlation Coefficient similarity values to distinguish between strong and weak social ties. These tie classifications are then used to personalize content recommendations and expose users to greater breadth or depth of applicable knowledge based on current interests or job goals. We simulate recommendations for a user under different circumstances and show that our recommender system promotes the algorithmic formation of communities of learners on similar educational tracks. This promotes the social-emotional support for online learners that they may not currently receive and improves socialization within distance learning.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2021
- Accession Number
- AD1150906
Entities
People
- Matthew F. Critchley
Organizations
- Naval Postgraduate School