Research And Implementation Of Collaborative Filtering Recommendation Algorithm Based On Integrated Optimization

With the continuous integration of the Internet into the social economy and daily life,people have become accustomed to various online information query and interactive behaviors(such as data retrieval,information query,online shopping,social networking,navigation,etc).The amount of network data and information and the rapid growth of its application has led to a research focus in recent years on how to accurately and quickly obtain the required information from the massive amount of information.Among them,various recommendation systems,as an effective method to solve this problem,especially collaborative filtering algorithms,have achieved many research results and have been widely used.However,related algorithms still have certain deficiencies and room for improvement in terms of sparsity,cold start,and resistance to shilling attacks.To this end,based on the development of related technologies,this paper conducts research and analysis on how to effectively solve the problem of shilling attacks and further improve the performance of the collaborative filtering algorithm,and proposes an integrated and optimized collaborative filtering algorithm to further improve the recommendation performance of the collaborative filtering algorithm.First of all,the research and analysis of the surrogate attack detection method based on the profile attribute and the time sequence feature are carried out.Based on this,a joint detection and filtering method of surrogate attack combining the above two methods is given and verified.It can realize effective detection and filtering of shilling attacks,convert polluted data sets with abnormal scores into pure data sets,and provide them to the collaborative filtering algorithm to improve the recommendation performance of the algorithm.Secondly,having studied and analyzed the research results of collaborative filtering algorithm,in addition,analyzed a variety of technologies that can be used for further optimization according to its insufficient research.Then,based on the "users clustering collaborative filtering recommendation algorithm combined with trust relationship",it is optimized and verified step by step,and an integrated optimized collaborative filtering algorithm is proposed.After verification,the algorithm can better solve the sparsity,cold start problems to improve the accuracy and completeness of project recommendations.Third,the algorithm in this paper is compared with traditional collaborative filtering algorithms and algorithms in other documents on the “Movielens” data set.The results show that the effect of the algorithm in this paper is more effective.Finally,so as to verify and apply the above research results,the design of the movie recommendation system is given and the functions of the system are implemented.