Bunda, Windi Reski (2025) Model Prediksi Titik Panas Kebakaran Hutan dan Lahan Di Kabupaten Kampar Menggunakan Algoritma Extreme Gradient Boosting. Other thesis, Universitas Islam Riau.

Text skripsi_203510056_watermark.pdf - Submitted Version Restricted to Registered users only Download (3MB) | Request a copy

Abstract

This research discusses the development of a forest and land fire hotspot prediction model in Kampar Regency, Riau, using the Extreme Gradient Boosting (XGBoost) algorithm. Forest fires in this area have become a recurring problem that negatively impacts the environment, public health, and the economy. This research aims to improve the accuracy of hotspot prediction by utilising MODIS and VIIRS satellite data and relevant meteorological data. The data analysis process includes handling missing values, normalisation, and dividing the dataset into training and testing sets. The XGBoost model was developed and evaluated, and compared with the AdaBoost model to measure the performance of each. The evaluation results show that XGBoost performs better than AdaBoost in all dataset sharing scenarios, with the highest R² value reaching 0.9328 after hyperparameter tuning. This research is expected to contribute to the government and related agencies in planning and mitigating the risk of forest fires, as well as increasing understanding of the factors that influence the occurrence of hotspots.

Actions (login required)

View Item