Using Collaborative Filtering Algorithms Combined with Doc2Vec for Movie Recommendation

Abstract

Information recommendation methods mainly include collaborative filtering and content-based. The collaborative filtering method is the most widely used recommendation method. It mainly uses the preferences of a group with similar interest or shared experience to recommend information of interest to users, but it will encounter serious data sparseness and cold start problems. In this paper, we propose a film recommendation model based on word vector features. The Doc2Vec model is used to extract the semantics, grammar and word order of the sentence, transform it into a fixed dimension vector, and the similarity of the vector will be calculated and applied to the collaborative filtering recommendation algorithm. Experiments show that the recommendation results are improved in both accuracy and recall.