Answers are grounded in this paper and cite their sources.
This study aimed to propose a model to predict the frequency of fire incidents and classify their severity in urban residential buildings, with a focus on fire prevention. Data on 92,000 fire incidents in London from 1981 to 2020 were collected. Data mining was then performed using the NumPy and Pandas libraries in Python to extract statistical information. The data were converted into numeric form using natural language processing (NLP) before modeling. In the modeling phase, a time series algorithm was utilized to predict fire frequency in London up to 2040, demonstrating that the fire frequency in 2040 would be 50% lower than in 2020. In the second phase, various machine learning models, including DT, RF, GB, SVM, and LR, were implemented for building fire severity classification. Among these models, the neural network demonstrated the highest performance, achieving an accuracy rate of 82%, outperforming the other approaches.
No takes yet. Share an insight, caveat, or question.
Behnam et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: