Research And Implementation Of Collaborative Filtering Recommendation Based On User

With the explosive growth of network information,it brings great difficulties to the users of use information.The problem of information overload can be solved by means of a kind of information filtering.At present,several kinds of recommendation algorithms are based on collaborative filtering,content-based recommendation and knowledge-based recommendation.The recommendation algorithm based on collaborative filtering is based on the user's rating of the product or the collection of the behavior model to provide users with personalized recommendation.There are several key steps in the recommendation method based on collaborative filtering.First,the calculation of similarity,similarity calculation can be different according to the different methods of calculation.The commonly used methods include Pearson correlation coefficient,cosine similarity method,Spearman method and so on.The second is to select the nearest neighbor of the user,the first choice should be filtered by threshold filtering,negative filtering,etc..Finally,this paper presents an improved recommendation strategy with cost factor in the system demand,the recommended strategy is introduced factor in the calculation of similarity to balance the impact of purchasing power for user similarity,so that similar user groups and the recommendation list are more suitable the user's purchasing power.After the collaborative filtering recommendation system based on user is implemented by web technology.Has achieved the management of wine products,user management,user collection management,front and back functionsthe.User login system and expressed interest in the system will be recommended for the wine might like.