深度学习耦合粒子群优化SVM的瓦斯浓度预测

Gas Emission Quantity Prediction Based on Particle Swarm Optimization of SVM and Deep Learning Network

为改进工作面煤矿瓦斯涌出浓度的预测精度,基于深度学习网络、SVM和粒子群(PSO)优化算法的原理,建立1种深度学习网络与粒子群优化SVM神经网络耦合的混合算法模型,该算法首先基于深度学习理论学习样本数据较深层次的特征,提取出较少个用来表征原始数据的特征量变量,对特征变量建立PSO-SVM预测模型进行瓦斯涌出浓度预测,通过工作面现场采集的数据进行仿真实验,实验结果表明该方法使预测精度较对原始数据直接进行PSO-SVM预测得到较大的提升,同时实现了原始数据的降维,减少了算法的运行时间,提高了算法效率。

In order to improve the accuracy and efficiency of coal and gas outburst prediction, based on the principle of deep learning network, SVM and the particle swarm optimization algorithm, the deep learning network was combined with particle swarm optimization of support vector machine neural network for the prediction of the situation of coal and gas outburst, more profound characteristics of variables can be learned from raw data through deep learning. Then the prediction models were established for a few characteristics by feature extraction which denotes the raw data by using PSO - SVM method to predict the emission concentration of gas instead of the raw data. Through mining workface of a certain coal mine in China to analyze and predict, the results showed that tile approach met the re- quirement that reduced dimensionality of the raw data, and the predict accuracy was improved greatly. Meanwhile, the running time and efficiency of the algorithm was improved through this method.