Work

Machine learning-based multi-model fusion for investment value protection

Year: 2025
Type: article
Abstract: Purpose: Extreme weather affects insurance risks and is highly uncertain. The protection of ancient buildings is an issue with prominent investment problems. This study aims to establish an assessment model to reduce risks in the insurance industry and raise the government's awareness of protection. Method: First, data from the United States and Australia from February 2020 to 2024 was collected and processed to solve problems such as multi-source heterogeneity. The ARIMA model was used to predict natural disaster data, and the parameters were determined after pre-processing. Based on EWM-TOPSIS to establish a regional risk assessment model to calculate weights and give underwriting strategies; based on the underwriting model to analyze building protection strategies, establish an ESC model, select indicators from multiple dimensions, use K-means and AHP hierarchical modelling to determine cluster categories, and calculate the comprehensive value score within the scope. Findings: Take Borobudur in Central Java, Indonesia as an example. The regional risk assessment model score is 32.6 (not recommended for insurance underwriting), and the building protection ESC model score is 6.65716 (the government needs to strengthen its protection). Research value: Insurance risk models help insurance companies manage risks. The government should pay attention to the value of scenic spots, provide effective strategies for building protection investment, and promote the scientific development of related work.
Authors: Haoyang Zhou, Luyu Ma, Hongyi Ling
Language: English
Open Access status: hybrid
OpenAlex ID: W4407602669
FWCI: 0.973
Cites: 6
Cited by: 1
Related to: 10