Network Personalized Recommendation Based On Multi-objective Immune Algorithm

With the rapid development of Internet and information technology,human have entered an era of data explosion.Facing with large amounts of information,we can't get useful information.The use ratio of the information reduces instead,which will make users confused on the information overload.So,how to find useful information from the huge amounts of data is imminent.Recommendation system,using the automatic statistics and data mining technology,is an effective means to solve this problem.It is able to provide users with useful recommendations,known as the most potential tool to deal with information overload.Usually,the main purpose of the traditional recommendation system is to maximize the accuracy.However,recommendation system only consider the accuracy of recommendation can't meet users.The personalized recommendation system give every user a recommendation list respectively,considering user's various needs.In the system,accuracy and diversity are two mutual restriction of performance indicators.Enhance the accuracy will be bad for diversity,and provide users different objects may also result in the decrease of the accuracy.Therefore,it is necessary to achieve an appropriate balance between recommendation accuracy and diversity simultaneously in the process of optimization.Immune optimization algorithm is an effective algorithm to solve the problem and provide a new thought by simulating the principle and function of the immune system.Immune optimization algorithm shows superior performance on global search and local search.In this paper,multi-objective immune algorithm is proposed based on the network personalized recommendation,by studying and exploring on recommend technology.The method uses the basic principle of biological immune system.The optimization problem was modeled as a multi-objective optimization to maximize the accuracy and diversity in the recommendation lists.Antibody coding and the immune operator were designed in the algorithm.Finally,the results show that the proposed algorithm can effectively obtain the optimal solutions of personalized recommendation,which improves the accuracy and diversity,so it recommend for multiple users with different needs at the same time.