Research on mobile service satisfaction prediction model based on GBDT regression algorithm

Abstract

With the advent of the 5G era, which has brought great convenience to people, people are increasingly inseparable from the various conveniences brought by various communication technologies. As network construction continues to evolve, mobile operators have begun to pay increasing attention to customer satisfaction with the network experience[1]. It is particularly important to improve customer satisfaction more effectively and comprehensively by analyzing the various factors that affect it[2]. GBDT is characterized by high prediction accuracy and is suitable for low-dimensional data. It can handle nonlinear data and can flexibly handle various types of data including continuous and discrete values. Moreover, GBDT can use some robust loss functions that are very robust to outliers, such as the Huber loss function and the Quantile loss function. Therefore, based on the Gradient Boosting Decision Tree (GBDT) model, this paper aims to analyze the factors affecting customer scoring, and predict customer scoring according to the factors.