Design And Implementation Of Movie Recommandation Of Movie Recommendation Engine Based Mahout

In the ever-changing tide of Internet entertainment, movies and other video sites offer video being unlimited number and variety of explosive growth, how users find the content they are interested in the mass of the video became a new problem. Personalized recommendation technology is emerging in this case, he may be interested in recommendation technologies recommended by digging favorite videos each user’s interests and hobbies, calculated by the machine for each user. For the current study abroad recommendation system is in full swing, for the purposes of personalized recommendation system, how to analyze historical information and user behavior accurately predict potential user interests and needs of its most core content, but also the focus of this study.This paper describes the first detailed study and include collaborative filtering recommendation algorithm based on association rules and content-based recommendation algorithm mainstream recommendation algorithm is widely used, in-depth study of the advantages and disadvantages of each recommendation algorithm Based on the above research and cutting-edge research at home and abroad for a hybrid recommendation system, a hybrid recommendation model incorporates content recommendation, association rules and collaborative filtering recommendation algorithm for multi-level switching.Then, the hybrid model is recommended, based on the design and implementation of a simple movie recommendation engine based on open source framework Apache Mahout, and describes in detail the engine the whole process from requirements analysis to architecture design to functional implementation. Meanwhile according to the authors’experience in large-scale systems development is proposed how to keep the current core of large video sites without changing the architecture and functionality, adding a recommendation engine module design ideas, making mutual decoupling between the base module and recommend module and reduce the risk of introducing the recommendation engine modules and cost.Finally, the paper design of hybrid recommendation models and movie recommendation system for analysis and assessment, explained that the system for widespread data sparsity and cold start defect in a certain degree of optimization, while noting the many problems which still exist, and on the future direction of the recommendation system put forward their views on major film class video site.