Personalized Recommendation Algorithm Based On Collaborative Filtering

With the increase of information quantity,users' score of data is more and more inaccurate,which is the problem caused by data sparsity.How to solve this problem efficiently has become the research hotspot of the current recommendation system.In this paper,according to the data sparsity problem existing in the existing recommendation system,we do more in-depth research.In order to improve the quality of the recommendation algorithm,this paper maximizes the use of user information,and constructs a joint similarity recommendation algorithm to improve the swarm optimization clustering integration.The steps of this algorithm are as follows: first,make corresponding modification to bee colony algorithm,and use the improved algorithm to process and optimize the center position of the algorithm in the cluster,so that the center in the cluster can not only get the best center in a certain range,but also the best center in the cluster within the range of all the data that can be used in this paper;secondly,the information between users in the data set and The user project information is integrated with the improved clustering algorithm respectively;thirdly,in order to solve the problem that the traditional algorithm uses one similarity algorithm,we use two different similarity algorithms to combine linearly to form a new similarity algorithm to calculate the similarity of different users in the same cluster;finally,the score of similar user information is larger than this The information of the average score of the user,which generates the recommendation list from high to low,is passed to the target user for selection.The experimental results show that this algorithm can not only alleviate data sparsity,but also improve the quality and accuracy of recommendation.The innovations of this paper are as follows:(1)For the data sparse phenomenon in the classical recommendation algorithm,this paper uses clustering algorithm to modify it deeply.By introducing artificial bee colony algorithm and integrated learning algorithm to optimize the clustering algorithm,several different data sets are used to verify the performance of the improved clustering algorithm,such as iris and red The data sets such as wine and movielens100 k are verified,and the results show that the algorithm in this paper is better than other comparison algorithms.(2)Aiming at the problem that the similarity of traditional recommendation algorithm is single,this paper uses two different similarity algorithms to calculate the information between users in the same category and the information between users' scores respectively,so as to form a new similarity,and make users' information get the most valuable use,and successfully implement the idea with the algorithm Now.