Movie Rating Prediction Recommendation Algorithm based on XGBoost-DNN

In the traditional movie recommendation, because the features of users and movies are not considered, only the users' ratings of movies are considered, so there is a problem that the recommendation is not accurate enough. In response to this problem, this paper proposes a movie rating prediction recommendation algorithm based on XGBoost-DNN. First, XG-Boost is used to screen user features and movie features, and the features that have a great impact on movie rating prediction are screened out, and then the screened features are used as the input of DNN, the user network, and the movie network is trained to obtain the user feature vector and movie feature vector respectively, and then the user's predicted rating of the movie is obtained through the neural network, and finally compared with LightGBM, SVR, KNN, and RandomForest, this paper proposed XGBoost-DNN model reduces the MSE indicator by 0.223, 0.75, 0.451, and 0.306 respectively, which effectively improves the accuracy of rating prediction, and thus improves the accuracy of movie recommendation.