<![CDATA[Forest fires pose a significant ecological and socio-economic threat globally, particularly in regions prone to climate change and anthropogenic pressures. This study conducts a comprehensive Systematic Literature Review (SLR) and Bibliometric Analysis of forest fire prediction models from 2021 to mid 2025, focusing on the utilization of machine learning, deep learning, and hybrid approaches. The review systematically analyses 65 relevant publications sourced from Scopus and ScienceDirect, selected through a PRISMA framework. The findings indicate that models such as 1D CNN, ConvLSTM, Transformer, and Random Forest are commonly applied, leveraging diverse datasets including satellite imagery, meteorological data, vegetation indices, and fire history records. Despite advances, most studies still emphasize meteorological variables, while local contexts such as socio-economic and land use factors remain underexplored. Furthermore, current models often face limitations in generalizability across regions. This study identifies key trends, gaps, and opportunities in the development of more robust and interpretable fire prediction models. This research also provides knowledge and insight related to the warning system for forest fire disasters. Recommendations include integrating socio-environmental data and developing geographically adaptive frameworks to enhance forest fire risk assessment and early warning systems.]]>
Copyrights © 2026
