Path Loss Prediction in Tropical Regions using Machine Learning Techniques: A Case Study

In optimization of wireless networks, path loss prediction is great importance for adequate planning and budgeting in communications. For efficient reliable communications the tropics, determination or estimation channel parameters becomes important. Research this article employed different machine learning techniques—AdaBoost, support vector regression (SVR), back propagation neural networks (BPNNs)—to construct models Akure metropolis, Ondo state, Nigeria. An experimental measurement campaign was conducted three broadcasting stations (Ondo State Radiovision Corporation (OSRC), Orange FM, FUTA FM) all situated within metropolis. Furthermore, we designed learning-based at various observation points a particular frequency, demonstrated how these algorithms agree with measured data. instance, OSRC (operating 96.5 MHz) measurement, RMSEs (root mean square errors) AdaBoost, SVR, BPNN, classical model (log-distance model) predictors were 4.15 dB, 6.22 6.75 1.41 respectively. Additionally, new frequency according to available data specific frequencies evaluated. order resolve challenge limited insufficient samples framework hybridizing developed. The developed employs estimated values that are computed by based on prior information training set expansion. Performance evaluation using FM (94.5 (93.1 MHz), from used as datasets frequency. log-distance 1.77 1.52 1.45 2.61 However, adding generated classical-based model, 1.81 1.63 1.88 results demonstrate proposed sample expansion enhances performance scenario few Finally, promising enough deployment technique practical scenarios.