Research on The Application of Video Recommendation Algorithm Based on DeepFM

Abstract

This paper studies the application of video recommendation algorithm based on DeepFM, studies the specific application of DeepFM algorithm in the field of video recommendation, and analyzes its performance. In this paper, the MovieLens video dataset is selected for experiments, and a simple video recommendation system is implemented by using UserCF, ItemCF and DeepFM. After recalling and sorting the dataset, the conclusion is drawn. The DeepFM model has superior performance in the sorting stage, and the loss is reduced by 0.55 % compared with the use of only FM model.