Research on geological hazard susceptibility evaluation using an information value-modified entropy weight combined model

Research on geological hazard susceptibility evaluation can effectively prevent threats to people’s lives and property from geological disasters. This paper selects eight influencing factors, namely, slope, altitude, distance to river, distance to road, precipitation, distance to fault, engineering geology, and the normalized difference vegetation index (NDVI), taking Qionghai city as the study area. By employing the modified entropy weight method, an information value-modified entropy weight combined model for geological hazard susceptibility evaluation is proposed. Susceptibility evaluations were conducted via the entropy weight method, modified entropy weight method, information value method, and combined model. The study area in Qionghai city was divided into four susceptibility zones: high, moderate, low, and very low. A comparative analysis of the results from the four methods revealed that the combined model yielded results that were most consistent with the actual situation. ROC curve analysis indicated that the combined model (AUC = 0.888) outperformed the entropy weight method (0.350), modified entropy weight method (0.830), and information value method (0.797) models, with significantly improved accuracy. Research has demonstrated that the modified entropy weight method overcomes the limitation of insufficient coverage of influencing factors during calculation. By constructing the information value-modified entropy weight combined zonation model, the problem of unsatisfactory evaluation performance when a single model is used is resolved. Simultaneously, key factors promoting and inhibiting the formation of geological hazard susceptibility zonation in the study region were identified. The results can provide a basis for government efforts in geological hazard prevention and control, as well as site selection and planning for major engineering projects.