Research On Customer Classification Of Securities Company Based On Neural Network

With the rapid development of securities market,the competition among securities companies has become increasingly fierce.There is a serious homogenization of services and products.The price war smoke everywhere.Only a few companies benefit and the rest in the loss outweighs the gain.Improving service quality is another way to get customers.Understanding customers is an important prerequisite for improving service quality,and reasonable classification of customers conducive to targeted work.This paper will study the customer classification based on F branch of A securities company.Firstly,the theories of customer classification and present situation of China's securities market are studied.Subsequently,the A securities company F branch's customer service is analyzed,of which the customer classification dimension and customer value evaluation index are too simple,lack of analysis and prediction of future value to the customer.Second,a customer classification method for F branch based on customer value is designed.The current value of customer is evaluated by using RFM model,and the future value is predicted by using a neural network model.The customers can be divided into four categories according to the current value and the future value:high current value and high future value,high current value and low future value,low current value and high future value,and low current value and low future value.BP neural network and RBF neural network are utilized to classify the customers of F branch.Simulation results show that the accuracy of BP is better than that of RBF.The accuracy of BP neural network in predicting customer's level of next quarter is about 78.27%,with the error of less than one level is less than 98.93%.BP neural network is more accurate than RBF neural network.Therefore,this paper selects BP neural network for customer classification prediction.Finally,the improvement of customer classification on marketing service quality is verified,as it helps effectively target customers and reduce workload,as well as advantageous to the targeted marketing services to improve success rate of marketing.This study provides a way for customer classification of securities companies.