A Movie Trailer Recommendation System Based on Pre-trained Vector of Relationship and Scenario Content Discovered from Plot Summaries and Social Media
Chun-Yu Chien, Guo-Hao Qiu, Wen‐Hsiang Lu · 2019
Posting articles on the social platform is the favorite activity of young people. With the potential of digital movie and tv series industry, developing an automatic movie recommendation engine becomes a popular issue. Traditionally movie recommendation is based on structured information like director, players, rough class, etc. Recently, there are more and more researches trying to make a recommendation based on context information like music recommendation based on lyrics with the word vector representation. However, in the long text scenario, the recommendation based on all context vector makes the inference very imprecise.In this paper, we propose effective features types, relationships, and scenarios, to extract important information then improve the recommendation. Furthermore, comparing different pre-training model, we try to maximize the effectiveness of semantic understanding and make the recommendation be able to reflect meticulous perception on the relationship between social media articles and movie plot.