MACHINE LEARNING-BASED INTERFERENCE DETECTION AND MANAGEMENT SYSTEM FOR 5G HETEROGENEOUS NETWORK

Interference remains a critical challenge in next-generation wireless networks, affecting both quality of service (QoS) and spectral efficiency. This study proposes a machine learning–based interference detection and management framework to address these challenges. A dataset consisting of key interference indicators including transmit power, path loss, antenna gain, user density, frequency reuse factor, and signal-to-interference-plus-noise ratio (SINR) was used to train and evaluate several machine learning algorithms. Artificial Neural Networks (ANN), Support Vector Machines (SVM), Logistic Regression, Decision Trees, and K-Nearest Neighbours (K-NN) were benchmarked using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and classification accuracy. Among these, the ANN achieved superior performance with a prediction accuracy of 99.78%, alongside the lowest error rates (MAE = 0.0022, RMSE = 0.0469), demonstrating strong generalization and robustness. To extend this predictive capability to real-time interference management, the study introduces a Dynamic Channel Sensing and Queuing Transmission Model (DCSQTM), which leverages Markov and Poisson processes for probabilistic estimation of channel availability and queuing delays. Machine learning predictions were integrated into DCSQTM to enable dynamic channel selection and adaptive transmission strategies. The combined framework significantly enhances interference detection and management, providing a scalable and intelligent solution to improve spectral efficiency and ensure reliable QoS in next-generation wireless networks.