COMPARISON LINEAR REGRESSION AND RANDOM FOREST MODELS FOR PREDICTION OF UNDERGROUND DROUGHT LEVELS IN FOREST FIRES

The increase in forest fires poses a significant risk due to its impact on underground dryness, which can cause long-term environmental damage and challenge fire suppression efforts. This research aims to develop a prediction model for underground drought levels in the context of forest fires using machine learning techniques. The methodology used in this research follows the CRISP-DM (Cross-Industry Standard Process for Data Mining) framework, which includes the stages of business understanding, data understanding, data preparation, modeling, evaluation, and deployment. This study analyzes a forest fire dataset, applies encoder labels to transform categorical variables, and uses linear regression and random forest models to predict underground drought levels. The goal is to create a predictive model that can help inform wildfire risk management strategies by anticipating underground drought levels. The results showed that the random forest model achieved higher prediction accuracy than the linear regression, with an R-squared value of 0.97. This suggests that the random forest model is a more robust tool for predicting underground drought levels, providing valuable insights for forest fire management. This research contributes to the understanding of underground drought levels, aiding the development of effective wildfire risk management strategies.