Maryam Munir

This study investigates deforestation trends in Rondônia, a critical region of the Brazilian
Amazon, by identifying key risk factors and predicting future deforestation using multivariate
logistic regression and machine learning methods. Logistic regression was applied to determine significant predictors, including roads, population density, conservation units, and agricultural activities. These predictors were then integrated into a Random Forest model to forecast deforestation scenarios. The findings highlight the impact of infrastructure development and farming on deforestation, while demonstrating the effectiveness of conservation units in mitigating forest loss. Future research should incorporate climatic variables and improve data access to refine predictive models.
Keywords: Deforestation, Amazon, Machine Learning, Random Forest, Logistic Regression,
Conservation Units, Infrastructure, Predictive Modeling, Rondônia, Spatial Analysis.