User Relationship Discovery Based on Telecom Data

With the improvement of the mobile communication market, the smart home business of telecom operators is showing unprecedented development and contribution. With the full-service competition in the domestic telecommunications industry, operators are facing increasing competitive pressure. The home service has become one of the important competitions for operators. In order to solve the user’s relationship recognition problem, this paper compares five commonly used machine learning methods: logistic regression, decision tree, random forest, LightGBM and multi-layer perceptron. Based the experiments on the data set, the results demonstrate the efficient of the LightGBM method. Finally, this paper selects the LightGBM method as the technical method for user relationship recognition in the business.