A Personalized Movie Recommendation System based on LSTM-CNN
Haili Wang, Na-Na Lou, Zhenlin Chao
Abstract
In the context of such an era where nearly everything is based on big data, personalized recommendation systems are becoming increasingly valuable for research. Deep learning has attained great achievements in numerous fields by virtue of its powerful computing power and extraordinary nonlinear transformation capabilities. Applying deep learning to a recommendation system that needs to mine and extract features from massive amounts of data will not only help the development of recommendation algorithms, but also improve the algorithm performance and thus improve the user experience. This project introduces a recommendation algorithm based on LSTM-CNN and applies it to the recommendation of movies by mining user behavior data and recommending movies with higher ratings to them. This article uses the data provided by the movie website MovieLens. It is testing set and training set that the data is divided into, and Top-N recommendation list is produced for the training set, while the algorithm is evaluated on the testing set. It is the features of the data that LSTM-CNN can effectively extract and complete the recommendation from the results.