Utilizing Explainable AI Methodologies: LIME and SHAP, for the Classification of Natural Disasters Through Machine Learning Algorithms
K Akshitha, R Roopashree, Ashwini Kodipalli, Trupthi Rao · 2024
This work investigates the use of machine learning for the classification of natural disasters, with a focus on the Random Forest and XG Boost algorithms with optimized hyperparameters. Explainable AI (XAI) techniques Lime and Shap provide insight into model selections, highlighting crucial characteristics such as ‘No Injured’ and ‘Total Damages’ that contribute to XG Boost's 72% accuracy. The impact of these features is graphically represented by the contribution graph. XG Boost outperforms other algorithms, demonstrating its effectiveness in classifying disasters. The study emphasizes how important XAI is for improving accuracy and offering perceptive judgments. In addition to improving model interpretability, it advances the subject of disaster classification and opens the door for other XAI applications.