Abstract
This paper proposes a risk evaluation model based on rough sets (RS) and the unascertained measure theory (UMT) for solving the accuracy problem of coal and gas outburst prediction with the aim to reduce economic losses and casualties in coal mining. The coal and gas outburst prediction problem is constrained by the selection of the prediction indexes, the coupling of a single index, and the weight of each index. The proposed RS-UMT model applies two modified techniques. The first one is a method for index weight determination that was improved by rough set theory. The second one is a method for coupling a single index that was modified by the unascertained measure theory. The RS-UMT model not only well solves the problem of coupling a single index of coal and gas outbursts, but also solves the problem that the weight is susceptible to subjective factors and prior knowledge. The RS-UMT model was used to judge the risk degree of outburst of 10 mining faces in the Pingdingshan No. 8 Mine and No. 10 Mine. The predictive results of the model were basically identical to the actual measured results. The performance of the RS-UMT model was also compared to existing methods. Based on the case study it can be concluded that the RS-UMT model is an accurate and very promising method for solving the coal and gas outburst prediction problem.