The Research On Customer Relationship Management Of Telecom Enterprise Based On Data Mining

With the fierce competition in telecommunication market and the development ofinformation technology, telecom enterprises’ competitive pressure increasing constantly.Telecom enterprises come to realize the importance of customer-centric, starting to changefrom business driven to customer-oriented. It’s a significant aspect for telecom enterprise toimprove the level of customer relationship management using data mining technology tomine and analyze a great deal of telecom enterprises’ customer data and discover various ofpotential, valuable and regularity knowledge. It has theoretical meaning and applied value.Aiming at two application problems of data mining in application of telecom enterprise’scustomer relationship management, the thesis takes the study with the method combiningtheoretical analysis and empirical research. The main research contents are as follows:1. Through a lot of literature investigation, the research focuses on analyzing two datamining application subjects of customer segmentation and customer churn prediction intelecom enterprise’s customer relationship management, pointing out the insufficient anddevelopment trend of current research, systematically introducing of customer relationshipmanagement theory, describing the general process of data mining and the main functionand algorithm of data mining.2. The paper establishes a customer segmentation model based on K-means algorithm,aiming at the specific case of a telecom enterprise, selecting attribute variable for customersegmentation, according to the enterprise’s customer transaction data, using SPSS statisticalsoftware for empirical study, depicting the characteristic of customer group of segmentationresult, and proposing corresponding marketing suggestions, at the same time, proposingsome application suggestions of segmentation model.3. The research establishes a model of telecom enterprise’s customer churn predictionbased on LVQ neural network, using of a telecom enterprise’s customer data, programmingwith MATLAB analysis software, learning and training of the neural network, and usingpart of high-value customer of customer segmentation results as the test sample set,assessing the result of the model, the fact shows that the model has relatively goodprediction accuracy, finally, proposing application recommendations of predictive models.The thesis provides telecom enterprise with beneficial reference in analyzing customerbehavior and advancing the level of customer relationship management using data mining technology, and shows a greater practical significance for telecom enterprise to maintain anddevelop customer relationship, enhancing enterprise’s competitiveness.