The Design And Implementation Of Insurance Recommendation System Based On Collaborative Filter Method

Information overload is an important problem in the Internet with application and users increasing hugely,and in that users cannot get the required information directly.Recommendation system can supply the information to users who want to get,which is a good method to information overload.There are many recommendation algorithms in the Internet applications and provide useful information to application users,which can be dived into recommendation based on contents,collaborative recommendation and hybrid recommendation.In recent years,the insurance business has undergone more and more changes.With the rapid development of the Internet,the insurance product also can be provided in the Internet in the form of virtual products.However,due to the wide variety of insurance,how to effectively meet customer's insurance needs is also becoming more and more important.Therefore,in the insurance product sales platform,in order to provide more information to customers correctly,a recommendation model based on collaborative filtering is proposed.Insurance product recommendation system includes four modules: user management,sales management,recommendation management and product management.The use case analysis,class diagram design,activity and sequence diagram analysis with UML(Uniform Model Language).The insurance sales platform users,products,user purchase records and user log information are analyzed.Based on the analysis of user portraits and product features,a collaborative filtering insurance product recommendation model is proposed.The recommendation m odel uses the user's portrait data to feature the product.The idea of "grouping by people,Classifying by goods" calculates the relationship between users,and recommends insurance products to users.In the recommendation system,one way is to cluster the users through K-Means clustering analysis algorithm,and recommend the products with the highest user evaluation to the users;the other way is to push the products often purchased in the group of friends and the insurance products with high evaluation through the reliable relationship between users and friends in the system.The insurance recommendation system includes a data layer,a data access layer,a recommendation layer and an application layer.The development and testing of the system was completed.The recommendation system helps users provide reasonable insurance products.Through the collaborative product recommendation system,which is based on collaborative filtering,the user's friend relationship data,the friend's product evaluation data and the user's browsing data,,and it greatly improves the user's satisfaction with the insurance product filtering and improves the insur ance Sales Performance.