ENVIRONMENTAL FRAGILITY BY MACHINE LEARNING ALGORITHMS

The advancement of predictive models by Machine Learning Algorithms (ML) associated with environmental data enables the improvement fragility, which are essential tools for decision-making. This study aimed to derive a prediction fragility testing ML covariates in state Minas Gerais. We use physical-environmental variables (soil, geology, climate, relief) weight attributes and calculation average obtain model Potential Environmental Fragility (PEF). Subsequently, we extracted PEF values 4,800-point grid, was used generate new called PEFML. based on five algorithms set 105 covariates. results indicated that best-performing PEFML Random Forest (R2 0.59 RMSE 0.47), indicating predominance low level. have strong correlations (0.7 Pearson); however, has stronger other data. Therefore, is robust captures information from coherent spatial patterns. Keywords: model; Spatial prediction; Forest; planning.