The Matrix Decomposition Based On Intelligent Optimization Algorithm In Recommendation System

Traditional recommendation system is a kind of Internet application system. Recommend to users commodities or services they might like or interested. In this era of information explosion, when people face a lot of information, it is difficult to filter out the information that really useful. Therefore, information filtering and efficient processing of information has become a problem of common concern. This has prompted the recommendation system came into being. Recommended system can proactively some interest or study user preferences, and then personalize the calculation, calculated by the system and find points of interest of users, Provided quickly and accurately recommendation for users, saving the their time and improve work efficiency.Matrix decomposition algorithm is an important algorithm in recommendation system. Some users of social media sites like to upload some photos to a place of tourism, these photos generally have geographic label, then these photos would provide people with a great deal of research data. Look these data as a matrix, using matrix decomposition techniques to achieve recommended. In this paper, we put forward a method of matrix decomposition based on particle swarm optimization and brain storm optimization algorithm, Use objective function minimization: To obtain travelers feature matrix U and landmarks feature matrixV. Use the Hidden features from U andV, we can use UTV to indicate the Likeability of travelers to landmarks. For the new data matrix, we arrange it according to their weights, produce the recommended list of Top_N, and recommend to travelers.We use hit ratio to measure the accuracy of recommend. The results show that the matrix decomposition based on particle swarm optimization and brain storm algorithm has the very big enhancement in accuracy and the algorithm efficiency, can acquire better global optimal solution, to improve the accuracy of recommendation.