Content Based Network Representational Learning for Movie Recommendation (CNMovieRec)

Nageswar Rao Kota, Vineet Padmanabhan, Wilson Naik Bhukya

Abstract: Research in the domain of recommender systems mainly points out to two types of recommendation strategies, namely, Collaborative Filtering (CF) and Content based filtering (CBF). In CF, the idea is to find users similar to the active user (the user to whom recommendation needs to be done) and recommendation is done based on items liked by similar users. On the other hand, content based filtering make use of the explicit or implicit data provided by the users on different items so as to generate a user profile and recommendation is made based on this profile. Previous research mainly focusses on CF-based approaches and many interesting results have come out in the recent years. In this paper we focus on CBF wherein a novel framework is proposed called content-based network representational learning for movie recommendation (CNMovieRec). We also propose a group recommender system based on our new CNMovieRec. Our experimental findings demonstrate that the results related to the proposed content-based framework is comparable to those of CF techniques (in terms of MAE, MSE & RMSE) and the new model is able to overcome challenges of Data Sparsity and Cold-Start problems usually accompanying CF-models.

External IDs:dblp:conf/miwai/KotaPB23