Explaining and Predicting Mobile Network Performance From Multi-Operator Data Using Machine Learning

This project focuses on the prediction and explanation of mobile network performance by leveraging machine learning techniques applied to multi-operator data. The application aims to provide insights into mobile network efficiency, including parameters like signal strength, data throughput, and network type, by employing advanced algorithms such as XGBoost, Random Forest, CNN, and LightGBM. The system facilitates the upload of mobile network performance data in CSV format, and users can register and log in to access personalized predictions regarding network performance metrics. Upon uploading, the dataset is analyzed, and performance metrics are displayed, including Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R² Score for the selected model. A prediction interface is also available, allowing users to input real-time data, which the system then processes through a pre-trained model to predict network latency. The application uses Flask for backend development, MySQL for data storage, and supports CSV file uploads for efficient data handling. By integrating machine learning models and interactive user interfaces, this project enhances mobile network management and helps operators optimize their services.

Keywords:

Mobile network, machine learning, prediction, XGBoost, Random Forest, CNN, LightGBM, network performance, signal strength, data throughput, Flask, MySQL, model evaluation, mobile operators, latency prediction.