瓦斯浓度作为衡量煤矿瓦斯危害程度的一个重要指标，为了能够更加准确的预测煤矿瓦斯的浓度，提出一种差分进化-分布估计（DE-EDA）算法优化的支持向量机瓦斯浓度预测新方法。利用无线传感网络系统采集工作环境中的瓦斯浓度数据，并经过降噪处理后作为训练样本。采用DE-EDA算法对SVM模型的惩罚参数C、损失参数ε以及径向基参数γ进行优化，利用优化后的模型进行瓦斯浓度的预测。通过MATLAB软件仿真可以得出，所采用的优化模型能够准确的预测煤矿瓦斯浓度的变化趋势。并与经过粒子群（PSO）算法优化的预测模型相比较。结果表明，经过DE-EDA算法优化的SVM模型具有训练速度更快、预测更准确的特点，为实际煤矿瓦斯浓度的预测和处理提供了更加可靠的理论基础。%As an important index to measure the degree of gas hazard in coal mine,in order to predict the coal mine gas concentration more accurately,proposes a new method for predicting gas concentration of support vector ma⁃chine based on differential evolution and estimation of distribution(DE-EDA)algorithm. The wireless sensor net⁃works system was used to collect the gas concentration data after the noise reduction in the working air. Used DE-EDA algorithm to optimize the parameters of the SVM model,penalty parameters C ,loss parametersεand radial ba⁃sis parametersγ,and predicted gas concentration by optimized model. Through the simulation of MATLAB software can get the conclusion that the optimized model can accurately predict the change trend of coal mine gas concentra⁃tion. Compared the conclusion with the prediction model of the particle swarm optimization(PSO)algorithm. The results show the SVM model has the characteristics of faster training speed and more accurate prediction,which pro⁃vides a more reliable theoretical basis for the prediction and treatment of gas concentration in coal mine.
展开▼
