This article establishes a customer rating prediction model based on customer online business experience data, aiming to help mobile operators better understand market operations and improve network service quality. Firstly, in the data preprocessing stage, missing and outlier values were removed, and data features were unified through feature engineering, including normalization, removal of irrelevant features, feature replacement, and encoding. Subsequently, the main factors affecting customer ratings, such as the total GPRS traffic (KB) accounting for $\mathbf{8. 7 9 2} {\%}$, were analyzed...
