| With the splitting and recombining of the domestic telecom providers and the issuing of the 3G licenses, the competition within communications industries and of wining the customers have become more and more tough, besides, the original content, quality, way and consciousness of service have been endured a severe challenge. DM technology, which helps gaining the advantage in a competitive customer center environment, is a large information container and an information processing tool. DM technology, which analyzes the extraction and depuration from a mass of information, aims at discovering the potential information. It is helpful for companies, especially in the following aspects: comprehensive apperception of customers, speedy discovery of the hidden business opportunities in abundance data and profound understanding of customers and their consuming behaviors, consuming favors and consuming values. As the result, the decision of a tailored package for clients will be set down through a suitable way in appropriate time.This article combines with communications industry operator's realities, analyses the customer consumer behavior with K-means clustering rules, C5.0 decision tree rules and apriori associate rules; apply to the customer consumer behavior in data analysis, designs and implements the customer consumer behavior analysis systems.In the first, this article introduces the related excavation data and development situation. Secondly, the needs analysis of the target and functional modules have been implemented in consumer behavior analysis system, while the system flowchart. The article uses clustering rules, decision tree rules and apriori associate rules to data mining and introduced the principles of clustering rules, decision tree rules and apriori associate rules detailed.In data mining module in the design, two clustering rules are used to compare. The K-means clustering rules is considered as the appropriate one can solve some of the attributes of data objects, issues together with local the end, the algorithms are relatively scalable and efficient, Order to enter data and sensitive algorithms are easy to understand the fruit, modeling and faster, and communications with the existing database characteristics, and they are K-means more applicable to consumption for the conduct of conclusion. This article takes the K-means clustering rules in communications provider customer segmentation for instance to expatiate the implement procedures of the algorithm, analyze the result, improve the algorithms, reduce the possibility of initial clustering the random selection and may occur in the local minimum out of convergence by K-means clustering rules, enhance the effect of the algorithms. Secondly, this system also uses the associated rules to data mining, taking whether the long packet and roam packet are suitable in bundle sales for example, expatiating on the associated rules apriori procedure, analyzing the result, improving the algorithms. Apriori associated with the rules for scans the database, and pattern match examination candidates set to low efficiency of the problems, improved apriori associated rules, only to scan a database, and greatly improve the efficiency of the algorithm. Third, the article introduces the relation and development process of the ID3,C4.5,C5.0,CART decision tree rules. It is thinking that the C5.0 decision rules are more applicable to the consumer behavior in the analysis. This system uses the C5.0 decision rules for marketing decisions of the filter. Taking the paper bag filter client for a long distance call for example, elaborating the implement process of C5.0 decision rules and analyzing the result.Practical applications, data mining techniques in the analysis of the consumer behavior can be achieved good results and assist enterprise to policy-makers insight into the consumer, thereby to increase corporate profits.Finally, this article summarizes the consumer behavior analysis system by data mart's。... |
PDF Full Text Request