With the advent of the 4G era, the competition in the telecom market is becoming more and more fierce, and customer resources have become the focus of competition among telecom enterprises. The law of customer consumption behavior is an important part of customer knowledge, so customer segmentation based on consumer behavior cognition has become the highlight of customer relationship management in telecom enterprises. Using data mining algorithm to analyze a specific customer consumption data set, we can dig out interesting information, and further adjust the marketing strategy of the enterprise according to these interesting conclusions.

In view of the shortcomings of current telecom enterprises in the segmentation of 4G customers, this paper realizes the segmentation of existing customers of telecom enterprises through correlation analysis, helps telecom enterprises to realize the reasonable classification of telecom customers, and thus puts forward guiding opinions on the marketing strategy of telecom enterprises. By analyzing an operator's 4G customer database, this paper uses Apriori algorithm to find interesting association rules between customer consumption behavior and consumption characteristics, and further analyzes this information, providing a new perspective for marketing decision makers.

The research idea of this paper is to preprocess the sample data, divide the sample data into three customer groups: changing 4G cards, changing 4G packages and changing 4G terminals, and then calculate the average monthly arpu value, average monthly mou value and average monthly dou value respectively. Finally, Clementine software is used to entropy group the three values of three customer groups based on MDLP principle, and the subdivided characteristic customer groups are obtained. Then do further research on these customer groups, use Apriori algorithm to generate frequent itemsets, generate simple association rules according to the frequent itemsets, dig out the association between customer consumption behavior and subdivision variables such as brand, arpu value, mou value and dou value, and summarize the corresponding laws to help telecom enterprises find the consumption habits of specific consumer groups, and then carry out targeted marketing on the identified consumer groups.