The mineable coal seam at Qidong coal mine, Huaibei coalfield, is adjacent to the overlying unconsolidated confined aquifer, which may induce water inrush hazards threatening the safe mining. This study presents an information value model (IVM) for the risk assessment of water inrush based on the information value and the geographic information system (GIS). In contrast to the general risk assessment model, this model was considered to improve the selection of the evaluation factor, which predicts the water inrush risk by using the best combination of factors. In general, the selection of the factors always had been decided by individual decision-makers. It was significantly subjective, which, in turn, make the prediction results quite different from the fact. Thus, it is of great significance to optimize the selection of evaluation factors before predicting the water inrush risk. This model was applied to the Qidong coal mine in North China. Then, the prediction results were further verified by the field distribution of water inrush, and the evaluation result is believed to be satisfactory. In addition, the AHP was also adopted for the risk assessment to compare with IVM. The AUC values of AHP and IVM without the optimization of factor selection are 0.6883 and 0.8229, respectively, and are 0.7253 and 0.8494 after the optimized selection, respectively. The results show that the IVM has a better prediction accuracy, and the prediction accuracy is increased after the optimization of factor selection, which indicates the IVM and the optimization could further help the future mining operation.

淮北煤田启东煤矿可采煤层与上覆松散承压含水层相邻，可能引发突水危害，危及安全开采。本研究提出了基于信息值和地理信息系统（GIS）的突水风险评估信息值模型（IVM）。与一般的风险评估模型相比，该模型被认为改进了评价因子的选择，利用因子的最佳组合来预测突水风险。一般来说，因素的选择总是由个体决策者决定。这是非常主观的，这反过来又使预测结果与事实大相径庭。因此，在预测突水风险之前优化评价因子的选择具有重要意义。该模型应用于华北启东煤矿。随后，通过突水现场分布进一步验证了预测结果，认为评价结果令人满意。此外，还采用层次分析法进行风险评估，以与 IVM 进行比较。未进行因子选择优化的AHP和IVM的AUC值分别为0.6883和0.8229，优化选择后的AUC值分别为0.7253和0.8494。结果表明，IVM具有更好的预测精度，并且经过因子选择优化后预测精度有所提高，表明IVM及其优化可以进一步帮助未来的挖矿作业。