Objective Evaluation of geological hazard susceptibility is an important means for prediction and early warning in the area of frequent geological hazard. Changyang County is located south of west Hubei Province and the middle and lower reaches of Qingjiang River,with frequent geological hazards. Therefore, high-precision susceptibility assessment is of great significance for disaster prevention and mitigation and for reducing disaster losses in this region.

Methods In this paper, 13 disaster−causing factors such as elevation, slope and undulation are extracted from four aspects: topography and geomorphology, geological environment, hydrological environment, and human activities. The I−ABC−LightGBM prediction model is constructed to evaluate the geological hazard susceptibility in Changyang County by using the informative model (I) coupled with the lightweight gradient lifting machine model (LightGBM), and ABC algorithm applied to optimize the parameters of the coupled model.

Results The evaluation results show that: ① the I−ABC−LightGBM model has high prediction accuracy, with a prediction accuracy value of AUC of 0.882, which is better than the 0.824 of the I model and the 0.851 of the I−LightGBM model; ② the extremely high susceptibility area and the high susceptibility area of the geologic hazards accounted for 32.37% of the total area of Changyang County, the medium susceptibility area accounted for 17.75%, the low susceptibility area accounts for 22.58%, the low−prone area accounts for 27.30%; ③ the very high and high prone areas are mainly in the middle and southwestern part of the Qingjiang River system along the coast and in the northern part of the Yichang−Wanzhou Railway and along the Hurong−West Expressway.

Conclusions This paper has methodological significance and typical examples of reference significance for similar geological hazard early warning, prevention and control, and land planning work.