关键词: 煤与瓦斯突出预测, 粗集, 粗神经网络, 混合系统, 属性约简

Abstract: A coal or gas outburst prediction model combining Rough-Set（RS） and BP Artificial Neural Network（ANN） is presented.RS theory is applied in analyzing coal or gas outburst dataset and the dependence relation between geological mining factor and coal or gas outburst is obtained on the basis of these data.So the feature elements are selected from the lager dimensions injections and regarded as ANN injection features，the number of the injection features can be reduced，and the “dimensions misfortune” problem caused by application of ANN coal or gas outburst prediction method to bulk power system is solved.The actual simulation example demonstrates that the model overcomes the disadvantages of constringency and has fast convergence speed and high prediction accuracy，compared with the single ANN method，and has an important practical meaning for the mine production safety.

Key words: coal or gas outburst prediction, rough set, rough neural network, hybrid systems, attribute reduction