This paper presents a research on operator user credit assessment using the LightGBM algorithm. The study aims to develop a credit evaluation model for mobile internet users based on massive data provided by communication operators. The process involves data preprocessing, feature engineering, and the construction of multiple machine learning models. Key features are extracted from the data to form a multi-dimensional feature set with statistical significance. Multiple basic models, including linear regression, decision tree, and LightGBM, are built to find the best model. Then, various integration algorithms such as Averaging, Voting, Blending, and Stacking are used to refine the models and establish the most suitable fusion model for user credit evaluation. The study highlights the advantages of the LightGBM algorithm in processing large datasets and its superior performance in credit assessment tasks. The research compares the performance of different models on four datasets: consumption capacity, location trajectory, application behavior preference, and others. The results show that LightGBM outperforms other models in terms of accuracy and performance. Additionally, the study introduces ensemble learning methods such as Voting, Blending, and Stacking to further improve the model's performance. The results indicate that the LightGBM-Stacking ensemble method is the most effective for credit evaluation. The paper also discusses the importance of feature engineering in improving the accuracy of credit assessment models. It emphasizes the need for a comprehensive approach that combines multiple machine learning techniques and ensemble learning to achieve better results. The study concludes that the proposed model provides a more accurate and efficient method for operator user credit assessment.This paper presents a research on operator user credit assessment using the LightGBM algorithm. The study aims to develop a credit evaluation model for mobile internet users based on massive data provided by communication operators. The process involves data preprocessing, feature engineering, and the construction of multiple machine learning models. Key features are extracted from the data to form a multi-dimensional feature set with statistical significance. Multiple basic models, including linear regression, decision tree, and LightGBM, are built to find the best model. Then, various integration algorithms such as Averaging, Voting, Blending, and Stacking are used to refine the models and establish the most suitable fusion model for user credit evaluation. The study highlights the advantages of the LightGBM algorithm in processing large datasets and its superior performance in credit assessment tasks. The research compares the performance of different models on four datasets: consumption capacity, location trajectory, application behavior preference, and others. The results show that LightGBM outperforms other models in terms of accuracy and performance. Additionally, the study introduces ensemble learning methods such as Voting, Blending, and Stacking to further improve the model's performance. The results indicate that the LightGBM-Stacking ensemble method is the most effective for credit evaluation. The paper also discusses the importance of feature engineering in improving the accuracy of credit assessment models. It emphasizes the need for a comprehensive approach that combines multiple machine learning techniques and ensemble learning to achieve better results. The study concludes that the proposed model provides a more accurate and efficient method for operator user credit assessment.