Lightning Prediction Model Based on Humidity, Pressure, Temperature and Wind Speed Dataset Using Machine Learning

Tan Kok Bin, Khairul Hamimah Abas*, Ruzairi Abdul Rahim

Abstract

Malaysia resulted in high number of lightning occurrence throughout the year due to its tropical climate and location. This paper presents the application of machine learning in predicting the occurrence of lightning events in KLIA region based on historical sensed lightning and meteorological data. Logistic regression and artificial neural network (ANN) are two prediction models utilized in this study to make a comparison in terms of accuracy of prediction. A three-layer neural network has been developed and the most suitable network with highest accuracy had been obtained after several testing of different structures. Meteorological data used in this paper are acquired from humidity sensor, pressure sensor, temperature sensor and wind speed sensor. These four variables are inputs to the prediction model. The output of the prediction model is a binary classification which indicates lightning occurrence or no lightning occurrence. The performance of the prediction model is evaluated using k-fold cross validation.