优化 DWT－SVM 参数并用于褐煤水分近红外检测

摘要

Orthogonal experimental design is a frequently used method to search the best experimental condition through calculates the response of variables.Usually , near infrared spectra model includes some parameters that influent the determination accuracy.Orthogonal experimental design was used to optimize discrete wavelet transform-support vector machines parameters when moisture in lignite coal samples was detected.The results showed that orthogonal experimental design can be used to effectively search optimal parameters with fewer experiments.%正交试验设计是一种常用于优化试验参数的方法，可以计算各参数对试验的影响。近红外光谱模型常常包含多个参数，参数设置直接影响模型检测精度。我们应用正交试验设计优化离散小波变换－支持向量机参数，对模型中包含的各参数选取七水平进行分析，并将模型用于褐煤样品的水分检测。研究结果表明正交试验设计可以通过较少试验次数有效优化模型参数。