Optimizing Wireless Network Performance with Decision Tree-Based Gain Prediction

This study investigates the optimization of wireless network performance using decision tree-based gain prediction. A simulation involving 100 users, uniformly distributed around a central Access Point (AP), calculates the Signal-to-Interference-plus-Noise Ratio (SINR) for each user, based on their distance and angle relative to the AP. The resulting dataset captures relationships between user positions and network performance metrics. A Decision Tree Regressor predicts continuous gain values, while a Decision Tree Classifier categorizes gains into discrete classes. The regression model demonstrates high accuracy with strong correlation, and the classification model achieves notable accuracy and recall, validated through confusion matrices and classification reports. Simulations evaluate the impact of both normal and predicted gains on network throughput and delay, showing that predicted gains closely approximate outcomes seen with normal gain calculations. These findings highlight the potential of decision tree models for optimizing network configurations and fine-tuning parameters to improve overall network performance.