<![CDATA[his study aims to analyze the characteristics of forest fires using the K-Means Clustering algorithm in RapidMiner software. Forest fires are disasters that significantly impact ecosystems and human life, making data-driven analysis of their causal patterns crucial. The dataset includes critical variables such as the Fire Weather Index (FWI) system components (FFMC, DMC, DC, ISI), weather conditions (temperature, humidity, wind speed, rainfall), and spatial coordinates from the Montesinho National Park in Portugal. The research methodology involved data preprocessing, feature normalization, and the implementation of the K-Means algorithm with three clusters to classify fires based on risk levels.The analysis revealed that Cluster 1 was dominated by high-temperature and low-humidity fires (high risk), Cluster 2 was characterized by higher rainfall (low risk), and Cluster 0 exhibited large-scale fires with significant wind influence. The clustering demonstrated the effectiveness of K-Means in identifying forest fire patterns based on environmental factors, supported by a Silhouette Score of 0.62, indicating reasonably well-separated clusters.These findings provide a foundation for developing more accurate early warning systems for forest fires and support data-driven prevention and mitigation strategies]]>
Copyrights © 2025
