Value Evaluation Twinning on Base Station Construction Based on Decision Tree and Logistic Regression

In order to fulfill the requirements of value- oriented wireless network planning, this paper proposes an intelligent algorithm to predict the value of base stations to solve the problem of inaccurate value evaluation caused by inadequate expression of features. The algorithm uses Gradient Boosting Decision Tree (GBDT) for feature fusion and Logistic Regression (LR) for high-value probability evaluation. Experiments are carried out according to the real network data, which proves that the algorithm is scientific and effective, with an accuracy of 95%, which could back up in prioritizing the construction of high-value base station and ensure the investment efficiency of network operators.