Research And System Design Of Personalized Movie Recommendation Based On Neural Collaborative Filtering

With the increasing number of online movies,people start to filter the information to get effective information,so personalized movie recommendation system becomes a hot research topic.In this project,we study the algorithm and application of personalized movie recommendation.The algorithm is based on the "user-item" pairs in the Movie Lens dataset,and the positive and negative samples are divided into positive and negative samples according to whether the user has selected the movie,and the training and testing sets are divided into positive and positive samples,and the effectiveness of the proposed algorithm is verified through comparative and parametric experiments.We also designed and implemented a personalized movie recommendation system based on an improved neural collaborative filtering model.Although collaborative filtering is widely used as an effective method for filtering information,the calculation of this algorithm relies too much on the scoring matrix,which faces problems such as matrix sparsity and scalability.Therefore,this topic introduces neural collaborative filtering,which optimizes the collaborative filtering algorithm by neural networkizing the process of matrix decomposition and avoiding the generation of scoring matrix.The neural matrix decomposition model is a specific implementation of neural collaborative filtering,and on the basis of this implementation,this topic proposes an improved model-a Neural Collaborative Filtering model fused with Bayesian Ranking.The model introduces factor decomposer and Bayesian personalized ranking,and the structure can further capture the nonlinear information of user features and optimize the ranking performance.Through experimental comparison,the new model has a certain improvement in hit rate and ranking performance compared with the original algorithm model,and maintains a fast convergence rate.Based on the neural collaborative fused with Bayesian ranking,this project implements a recommendation engine based on this algorithm and builds a personalized movie recommendation system based on B/S architecture.The system can collect user behavior,calculate algorithm and return recommendation results through database storage,and realize various business functions such as user personal center,operator page,movie detail page,system homepage and movie recommendation,etc.It provides a system implementation scheme of neural collaborative filtering algorithm.