Design and Implementation of Handoff Decision Making Algorithm using Machine Learning Technique

Abstract

This paper focuses on the design and implementation of handoff decision making algorithm using machine learning technique. In wireless communication, mobile operators and their subscribers always experience dropped calls as a result of technical error in handover process which affects the quality of service (QoS) in the network. The technique employed used decision tree, random forest, gradient boosting and multi-layer perceptron as machine learning models which were trained to learn patterns that map feature space to handoff decisions. The results obtained indicates that at low mobility band, the decision tree produced 99.22%, 98.12%, 98.29% and 98.20% as an evaluation metrics for accuracy, precision, recall and F1-score. Also, at medium mobility band, the gradient boosting produced 97.91%, 95.38%, 94.92% and 95.14% for accuracy, precision, recall and F1-score. Again, at high mobility band, the multi-layer perceptron gave 99.49%, 98.44%, 99.49% and 98.96% for the same evaluation metrics. This shows that at different mobility band, the values of the evaluation metric vary depending on the type of mobility band. The machine learning technique is recommended for the applications of handoff decision algorithm and future generations of network.