Analysis Of Customer Loss Of Mobile Commuication Operators Based On BP Neural Network

With the more and more extensive application of artificial intelligence in society,how to effectively use artificial intelligence to analyze customer churn has become the focus of researchers.According to relevant research,for operators,the cost of redeveloping a new user is six times that of the cost of retaining an old user.In addition,the profits of the old users are 16 times that of the new users,so how to effectively reduce customer churn is very important for mobile operators.In this paper,we mainly use the BP neural network model to predict the loss of mobile operators' customers.Firstly,according to the situation of feature selection in customer churn data of mobile operators,this paper uses K-means clustering to classify customer groups into five categories,and screens out the feature variables of each category by drawing probability density diagrams of lost users and non-lost users.However,due to the imbalance between positive and negative samples in the sample data,the number of positive and negative samples tends to be the same after the sample is processed by smoteenn method.As the problem of customer churn of mobile operators is a high-dimensional,complex and non-linear problem,the model of BP neural network is used to analyze the problem of customer churn of mobile operators.The empirical analysis results show that the classification method has achieved good results in customer churn prediction,and its accuracy can reach 80%.Finally,using the characteristics of Deep Neural Network(DNN)multi-hidden layer and reverse auto-adjusting weights and bias values,we predict the loss of various types of consumer customers,and find that the accuracy can reach 92% or more.