中文摘要：

为了对煤与瓦斯突出进行有效预测,将遗传算法和支持向量机相结合,提出煤与瓦斯突出预测的GA-SVM模型.以我国典型的煤与瓦斯突出煤矿15个实例为样本,以交叉验证准确率作为遗传算法的适应度函数,搜索得到径向基核SVM最优惩罚因子C=28.8786、宽度函数σ=0.16508,利用最优参数建立煤与瓦斯突出预测GA-SVM模型进行预测,结果与实际完全一致.应用该模型对云南恩洪煤矿8个突出实例进行预测,并与单项指标法、综合指标法和BP神经网络进行比较.研究结果表明,煤与瓦斯突出预测GA-SVM模型具有较高的可靠性和精确性,能对煤与瓦斯突出进行有效预测.

英文摘要：

In order to predict coal and gas outburst effectively, this paper which combined with genetic algorithm（GA） and supported vector machine（ SVM）, established a coal and gas outburst prediction GA- SVM model. 15 typical examples of coal and gas outburst coal were collected to es- tablish GA - SVM model. The fitness function of genetic algorithm were established by the cross validation accuracy of the samples, and the final search best results of penalty factor C = 28. 8786 and the width function of RBF a = 0.16508. Using the optimal parameters to establish coal and gas outburst prediction GA - SVM model, and accuracy of the model was perfect. To further verify the accuracy of GA - SVM model, coal and gas outburst instances of Yunnan Enhong mine was used to verify this model, showing that GA - SVM model had an excellent performance and a high prediction accuracy. The results showed GA - SVM model had a high credibility in assessing coal and gas outburst, so it was a new approach to forecast the coal and gas outburst, which could be applied to practical engineering.