Research And Implementation Of Stock Dividend Distribution System Upgrade Programme In H Securities Company

With the rapid development of the securities industry,the number of customers and transaction volume of H securities company is growing rapidly,and many functions of current securities broking services system is fall behind business and market requirements.The working flow of stock dividend distribution system in H securities company involves many manual work,some functions are outdated and no longer fulfill market requirements.To solve these problems,current system needs to be upgraded to improve working efficiency,reduce costs and improve customer satisfaction.This article explores the upgrade solution of the stock dividend distribution system of H securities company.Firstly,we analysis the functional modules of existing system,and communicate with business users to identify the problems of their root causes.Secondly upgrade solutions are worked out against existing problems.The upgrade solutions mainly focus on four modules which are standing data maintenance module,holding adjustment and election module,transaction module and enquiry & analysis module.After system upgrade,corporate event information will be captured automatically,corporate event election will be extended to front end web trading platform,the data between front end and backend will be synchronized,election by standing instruction and election by neural networks will be introduced,the data mining algorithm will be introduced for stocks clustering and to analyze customer values.During system design stage,the overall upgrade solution is made according to the requirements of stock dividend distribution,detail design including database design,user interface and system functional modules design are completed.In system testing stage,detailed testing including functional test and performance test are conducted.The testing result is positive which indicates all functional modules meets user requirements are the system is ready for go live.Considering the high performance,high availability and high extendibility requirements for securities system,we choose IBM AS/400 as the platform of backend system with integrated DB2/400 as database,and Windows Server in front end with database Oracle.IBM Web Sphere MQ is used for real time data transmission between front end and backend.Historical election data is being trained by BP neural networks for a classifier for corporate event election,which is a new application of neural networks.K-Means algorithm is used for stocks clustering base on annualized dividend rate and we enhanced the RFM model to analyze customer values.