Leveraging Radial Basis Function Neural Networks with Adaptive Variability Functionality to Predict Wild Fires

• Published in: 2023 4th International Conference on Smart Electronics and Communication (ICOSEC) pp. 1583 - 1588
• Main Authors: Krishna, T V Sai, Bhargavi, V, Jyothsna, S, Rakesh, Vuduga, Bikshapathi, Dadi, Saritha, D
• Format: Conference Proceeding
• Language: English
• Published: IEEE 20.09.2023
• Subjects: Adaptation models; Adaptive systems; Biological system modeling; Classification; Climatic Changes; Forest fires; Forestry; Planets; Radial basis function networks; Severe Heat Waves; Weather forecasting; Wildfires; Wind Pattern
• DOI: 10.1109/ICOSEC58147.2023.10276328

A forest fire is a risk to the environment that has an impact on both human and animal life. So it's crucial to forecast forest fires. Nowadays, a lot of methods for predicting forest fires are based on Machine Learning (ML) and Deep Learning (DL) algorithms. By delivering solutions to numerous crises, like forest fires, ML is assisting the planet. In this work, a Radial Basis Function Neural Network with Adaptive Variability Functionality (RBFNN-AVF) was proposed. To anticipate forest fires in images, the ML method RBFNN-AVF is used in this work. Building a model for early wildfire identification and aiding in damage management due to such occurrences are the goals of employing such a system. There are several approaches to finding forest fires, including sensor detection and real-time geological data processing. However, utilizing picture classification, where ML is the most efficient method, is one of the greatest techniques to detect fire. ML techniques allow for the addition of these response systems or the configuration of drones with them, allowing for the frequent taking of pictures from the sky, the detection of smoke in dense forests, and the prompt notification of the appropriate authorities. Using the RBFNN-AVF approach, \ fire detection is carried out using the suggested model, and the results were successful in terms of accuracy. Therefore, using a model based on ML to handle calamitous scenarios may be an option. Out of all of these classification models, RBFNN-AVF had the highest precision of 96.9%. According to the results, it is possible to predict a wildfire's size before it even starts by analyzing climate data.