Computation of VHF Signal Strength for Point to Area Network using Machine Learning Modeling Techniques

Abstract

In this paper, computation of very high frequency (VHF) signal strength for point to area network was carried out using machine learning modeling techniques. Seven different machine learning models were adopted: Decision Tree, Random Forest, AdaBoost, k-Nearest Neighbor, Support Vector Machine, Artificial Neural Network and Linear Regression. A total of 120 data points was used in computing the signal strength. 72 data points (60%) was used to train the model, while the remaining 48 data points (40%) were used as test data to determine the accuracy of the computation for all the models. From the results, it was observed that the accuracy of the computations was greatly influenced by the amount of training data that was used. Also, from the results, in highest order of accuracy, AdaBoost was adjudged the best model. This was followed by the Artificial Neural Network model. Generally, the error margin of computation obtained for these two models were low, hence indicating that the models can be effectively relied on for computation of signal strength in the study area.