Recommendation Strategy Research And Application Of Small And Medium Sized B2C Cosmetics Mall Based On Customer Segmentation

With the development of economic globalization,Internet has brought more broad market prospects for the development of SMEs.In view of a large number of female customers in online shopping market, this paper designs and realizes small and medium sized B2C mall system based on cosmetics sales,to provide customers with an online shopping platform,to help enterprises better profitability and development.In the implementation process of the system,to improve the user experience as the standard,optimize the system performance.And through the analysis of customer behavior and value,and puts forward a customer segmentation method based on HSDPSOK algorithm and the method to realize the customer classification,personalized commodity recommendation,mining potential customers and for precision marketing has important significance.The main research contents of this paper are divided into three aspects,as follows:Firstly,this paper according to the specific needs of the business enterprise,introduces the Struts2,spring and Hibernate three framework structure,process and characteristics,detailing integration framework SSH2,cosmetics store system is developed based on the framework design.Secondly,this paper designs and implements the following three strategies to optimize the performance of the system.(1)Through the establishment of appropriate index and introducing ehcache cache to improve the efficiency of data query,implement database performance optimization.(2)The paper presents intelligent caching strategy based PV statistics,to achieve thesystem previously intelligent retrieved the datamay be accessed by the client from the database and put them into the cache,thereby increasing the search efficiency of the system.(3)Through the realization of paging to improve the efficiency of the query results show,reduce customer waiting time.Finally,this paper proposes a customer segmentation method based on HSDPSOK algorithm.Customer segmentation using clustering algorithm,K-means algorithm is widely used,but there is a need to pre determine the number of clusters of K and easy to fall into the local optimum of the two major defects.In order to solve these two problems, this paper introduces a SD algorithm and PSO algorithm,and proposes a HSDPSOK algorithm to improve the quality of clustering:(1)Clustering quality evaluation by SD to solve the problem of clustering number K. Using the sample data corresponding to different K values for the initial K-means cluster,using the validity index for comparison,return the K value that corresponds to the SD minimum,to ensure that the use of the best K value for clustering;(2)By using the PSO algorithm to solve the initial cluster center selection problem.Based on particle swarm optimization algorithm global optimization capability,select the optimal initial point of K,as the initial cluster centers.Experiments on the customer data set and analysis of the experimental results,through the evaluation indicators,it is proved that the use of HSDPSOK algorithm for customer segmentation can achieve better results.Finally,thepaper recommends goods based on the classification results achieved by the algorithm.