基于GWO的SVM在红外甲烷传感器测量误差分析中的应用

针对传统支持向量机回归模型应用在红外甲烷传感器测量数据处理时出现预测精度低的问题,提出了一种基于灰狼优化算法的支持向量机回归模型。该模型在传统支持向量机的基础上,利用灰狼优化算法自适应搜索特征空间来选择最佳特征组合,经过循环比较,能快速、准确地搜索到最优的惩罚因子C与gamma参数。用实验室研制的红外甲烷传感器对0～5.05%浓度范围的标准甲烷气体进行测量后,建立了3种SVM回归模型,并进行对比。结果表明,采用灰狼优化算法建立的支持向量机回归模型其绝对误差和相对误差小,精度高。

Application of SVM Based on Grey Wolf Optimizer in Measurement Error Analysis of Infrared Methane Sensor

A SVM regression model based on grey wolf optimization (GWO) algorithm was proposed to solve the problem of low prediction accuracy when the traditional support vector machine (SVM) regression model was applied to measurement data processing of infrared methane sensor. Based on the traditional support vector machine, the model used the grey wolf optimization algorithm to adaptively search the feature space to select the best feature combination. After cyclic comparison, the model could quickly and accurately search for the optimal penalty factor C and gamma parameters. After the measurement of standard methane gas in the concentration range of 0～5.05% with the infrared methane sensor developed in the laboratory, three SVM regression models were established and compared. The results showed that the support vector machine regression model established by the grey wolf optimization algorithm had the smaller absolute and relative errors and the higher accuracy.