Research And Implementation Of Personalized Movie Recommendation System Based On Web

Currently in a network of information explosion era, network information overload has become an urgent problem, the recommended system is an effective means to solve the information overload, by recommendation system can help users quickly discover potentially valuable content. The traditional model is recommended to run the recommended system deployed on a single machine, but due to the current number of users and the amount of network data to be recommended contents enormous, beyond the processing capabilities of traditional recommendation system, it is necessary to study the recommendation system based on distributed platforms Hadoop.By using recommendation system based on Hadoop platform to solve the traditional recommendation systems present problems, But at the same recommendation system based on Hadoop normally present data sparsity,cold start,the recommendation result lack of novelty and other issues.This paper analyzes the causes of these problems, the design of efficient and scalable tiered hybrid recommendation model as the focus of the study. Combined with Distributed File System HDFS and Map Reduce programming model, Map Reduce programming model is designed based on a hierarchical distributed parallel hybrid recommendation algorithm realization, and on this basis to achieve the recommended system prototype based on Hadoop platform. The main contents are as follows:1. Research on content-based and collaborative filtering recommendation algorithm specific process(both technologies), analysis of the advantages and disadvantages of these two recommended techniques, and these two algorithms were improved adaptive. Generate data sparsity for collaborative filtering algorithms, cold start, and content-based recommendation of lack of novelty, the researchers designed a hybrid recommender system based on hierarchical these two algorithms to solve these problems exist recommendation systems.2. To study the Hadoop operating mechanism, the analysis of specific processes Distributed File System HDFS and Map Reduce programming model. Combined with hybrid recommendation algorithm design and implementation of Map Reduce distributed hybrid recommendation algorithm parallelization scheme, and hybrid recommendation algorithm for Map Reduce workflow optimization. Implement a Hadoop-based recommendation system prototype based on this system through somebasic functionality testing, and stable operation.3. Binding recommendation based on collaborative filtering and content, presents a layered hybrid recommendation model to improve quality recommend recommendation system. And on this basis, design and implementation of the recommendation system prototype. Key design prototype system design phase of collaborative filtering and user-based collaborative filtering algorithm of Map Reduce-based items.4.Studied Hadoop, mahout and other open source software, combined with some clustering algorithms mahout, and achieve the recommended algorithms. Installation and deployment of these open source software, improve the recommendation system prototype based on the open-source algorithm mahout.