A Hybrid Movie Recommendation Algorithm Based on Optimized K-means Clustering
Zhixiang Lu · 2022 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · 2022
With the advent of the big data era, there is an increasing amount of information available on the internet, resulting in an explosion of information and information overload, making it impossible for users to access the information they want in a short period of time. In this regard, the recommendation system uses algorithmic deep mining of users' personal information and historical data to find out what they want and gives them a list of personal recommendations. The collaborative filtering algorithm looks for users with similar behaviour, the algorithm is not affected by new items and has a better recommendation accuracy in comparison. This paper analyses and compares content-based and user-based collaborative filtering algorithms and adds clustering algorithms to the algorithm's practical optimization process. The collaborative filtering algorithm suffers from poor recommendation accuracy, so the time-based interest calculation is added to optimize the similarity formula. Before calculating user similarity, the clustering algorithm is used to cluster users, aggregating users with similar behaviour interests into a class. Then PCA algorithm is used to reduce the dimension of project attributes, and finally the similarity was calculated by combining the optimized user data to obtain the final Top-N recommendation list. The experimental data were obtained using the public Movielens dataset, with 70% of the data used in the training set and 30% in the test set. The experiments show that the model is effective in extracting user features, while the recall and accuracy of the inference and storage techniques is improved. The hybrid optimization of the kmeans model with collaborative filtering is therefore a great contribution to the development of recommendation systems.