Decision Tree Analysis On Consumption Behaviors Of Unicom Customers In Tianjin

With the continuous improvement of technology,the data of user behavior mastered by the communication industry is more comprehensive and complex.At the same time,with the development of big data field,the application of data analysis in communication industry is imperative.At present,the most urgent application scenario is the war between competing operators for customers,among which the classification prediction of the customer is the focus of the operators.In the Unicom industry,broadband business is the strategic foundation of China Unicom,and it is the foundation and core to promote the sustained and healthy development of the whole business.How to effectively classify and predict the broadband customers,how to analyze the potential attributes of all types of customers,and how to formulate appropriate marketing strategies are the most concern of operators.The tool to solve the above problems is the data mining,which involves a variety of techniques and algorithms.As an important method in data mining,decision tree algorithm development will also face more content and challenges.Therefore,based on the intelligent data mining,it establishes the opening wisdom wojia sharing effect of different packages to analyze the decision tree of the Unicom version calls,SMS charges and value-added fees.It also divides the customers in details and studies the corresponding theory research to provide reasonable suggestions for the enterprise.The main content is as follows:(1)To respectively study the effects of customers' age,gender and the kind of the package on city calls,SMS charges and value-added fee,and make a crosstabs analysis on them.According to the confidence,the paper analyzes whether each genus has a significant impact on city calls,SMS charges and value-added costs.(2)Using decision tree CHAID algorithm,a decision tree,Unicom intelligent home sharing customer decision tree analysis is established.On the basis of the establishment of customer decision tree,it analyzes the impact's degrees of the natural attributes and package types of most customers who open different packages with wisdom wojia shared version on city calls,SMS charges and value-added fee.And then according to the characteristics,it makes a targeted marketing,thus reducing the cost of broadening the business market for the company.Finally,the validity of decision tree is established.According to the confidence interval and standard error,it determines the validity of the decision tree.Through the analysis,the age of the customers has the largest effect on the local telephone fee,followed by package type,and gender comes last;in the SMS charges,only package type occupies the largest influence,that is to say the package messages can meet the needs of the vast majority of customers with few people having SMS fee;in value-added fee,the package type also has the largest effect on it,followed by sex.The bigger package capacity the customers have,the more value-added service fee will be.As the communication enterprises,they should focus on understanding the consumers' behavior characteristics,their customers' satisfaction and their usage,and therefore,they can make a targeted marketing,which can greatly reduce the cost of enterprises in order to expand the business market.