Improvement Of Collaborative Filtering Recommendation Algorithm

With the explosion of information and data on the growth of the Internet,users need intelligent access to information in a manner changing,experience the evolution of portals,search engines and recommendation system.Recommended system is also recommended by the previous development of the current popular personalized recommendation system.The so-called personalized recommendation refers to the recommendation,while ensuring different user access to information in line with their own personality,their intelligence has made substantial progress compared to the previous information acquisition mode,also has a better user experience.Collaborative filtering technology is personalized recommendation technology is the most successful and most widely used technique that by calculating the similarity between users to select the target user's nearest neighbor set,and then by scoring nearest neighbor set to predict the target users Unrated The project score,resulting in recommendations.However,since the user program ratings extreme sparse matrix and other reasons,collaborative filtering recommendation precision technology still has much room for improvement.From collaborative filtering algorithm to improve the precision in recommending starting to do the following two things work.To solve the privacy security problem of the recommendation algorithm between systems,This paper developed a secure computation model based on the theory of secure multi-party computation.The model used LBlock,a lightweight block cipher algorithm,to encrypt the data provided by the third part,and RSA public key cryptosystem to manage keys of LBlock.Applying this model to the collaborative filtering between systems with randomized perturbation techniques,the paper developed a new algorithm whose calculation method of similarity can protect the system from the attack of artificial users.Secure vector was used to prevent the untrusted third party from colluding.Experiments show that algorithm not only has stronger ability to protect the user' s privacy disclosing to the system which is cooperated,but also has better quality of recommendation.Considering the effect of the surprise score,the popularity of the project,and the high quality high quality on the similarity computation of traditional collaborative filtering algorithm,a new collaborative filtering algorithm based on surprise and reducing popularity is proposed.In the process of similarity calculation,the author weighs the advantages and disadvantages between the surprise score and the accuracy of the recommendation to get a better set of neighbor.In order to evaluate the surprise score of the recommendation result,An evaluation rule which calculates the distance from the recommendation result to user.Experiments show that algorithm not only has stronger ability to predict the score of the target,but also makes the recommendation result a high surprise score for the target user.