APPLYING COLLABORATIVE FILTERING METHOD FOR DOCUMENT RECOMMENDER SYSTEM

Abstract

The recommender system helps recommend relevant information items to the user. In recommender systems, collaborative filtering is commonly used to gauge users' interest in new products. Collaborative filtering systems often rely on data about the similarity of users or products in the system in the past to predict preferences or new products for specific users. In this article, we apply the collaborative filtering technique with the k-nearest neighbor to recommend documents for the English center. The implementation process includes the following steps: Firstly, we build a system to collect and store data in the database; Secondly, we implement a recommendation algorithm with three cases, including Case 1 for new users, Case 2 for users who have seen the most document items, and Case 3 for centers' members. The results make it easier for users to find documents.