Mobile User Satisfaction Analysis Based on Decision Tree and Random Forest Algorithm

With the popularization and development of mobile communication technology, users of mobile communication services have seen a massive growth, and customer satisfaction, as an important indicator of operator services, is analyzed to help operators improve their service quality. In this paper, we use decision tree and random forest algorithms to analyze customer satisfaction with respect to the given data. We first performed data pre-processing, after which this paper introduced decision tree classification model and regression forest prediction model to analyze the data, derived the characteristic importance of each variable, and finally calculated the weights of the factors affecting customer satisfaction. After observing the difference of variables in the original data set, we obtained the final influencing factors by taking the intersection, and thus made the random forest prediction of customer satisfaction. The final classification prediction results were tested to ensure the accuracy was greater than 95%. We also explain the rationality of the model.