Research And Application Of Data Mining Technique In The Enterprise CRM System

Traditional marketing model has been challenged by the impact of rapid development ofinternet and increasing competition between enterprises, which has made it necessary forenterprises to apply advanced management idea and technology in order to fully acquire andmaintain customer information, discover potential market opportunities, avoid risks, and increasecustomer loyalty and satisfaction. The ability of recording as well as utilizing customer and marketdata has become a crucial factor that may affect enterprises success, accordingly there must be aeffective system to select analyze and utilize customer information. CRM (Customer RelationshipManagement) came into page aiming to assess enterprises quality and improve theircompetitiveness by processing customer management and resource integration, thus making themstand in an invincible position in the current competitive environment. At the same time, alongwith the expanding of enterprise businesses, abundance of the whole data informationexponentially grows, including the valuable massages and the redundant data, which makes itdifficult to extract expected resources from massive information then chart client network. Inconclusion, construction of a complete CRM system needs the combination of traditional CRMand latest data mining techniques.This paper mainly discussed the CRM system in China Tobacco Zhejiang Industrial Co.,LTD(CTI), demonstrated their data analysis method, categorize customers thus chart client networkand test customer loyalty. The CRM system plays a vital role in CTI development. Our studyfocused on4aspects below:1) Analyze CTI’s current situation, considering the advantages, framework and operatingmechanism of CRM system, and assess the feasibility of CRM system and viable use of datamining techniques according to enterprise’s demand.2) Draw a customers classification tree diagram to show their behavior and imply theirloyaltystage.buildthealarmingmodel so thatwe cananalyzecustomer’s loyaltyin specificperiodor at a specific time point thus giving possible solution to CLV curve.3) Demonstrate a new method to extract client network chart based on modified FP-Growthalgorithm, therefore decision makers are able to find out the relationship between customers andmake a reaction in time in order to improve client relationships. 4) Put forward a novel text mining technique based on TFIDF algorithm, which can reducethe service stress and make it more efficient to extract expected information so that satisfycustomers.