基于小波核偏最小二乘回归方法的混沌系统建模研究

基于核学习的强大非线性映射能力,结合用于回归建模的线性偏最小二乘(PLS)算法,提出一种小波核偏最小二乘(WKPLS)回归方法. 该方法基于支持向量机使用的经典核函数技巧,将输入映射到高维非线性的特征空间,在特征空间中,构造线性的PLS回归模型. PLS方法利用输入与输出变量之间的协方差信息提取潜在特征,而可允许的小波核函数具有近似正交以及适用于信号局部分析的特性. 因此,结合它们优点的WKPLS方法显示了更好的非线性建模性能. 将WKPLS方法应用在非线性混沌动力系统建模上,并与基于高斯核的核偏最小二乘 关键词： 小波核 偏最小二乘回归 混沌系统 建模     

Uniform design and regression analysis of LPCVD boron carbide from BCl3-CH4-H2 system

Boron carbide was prepared by low pressure chemical vapor deposition (LPCVD) from BCl₃-CH₄-H₂ system. The deposition process conditions were optimized through using a uniform design method and regression analysis. The regression model of the deposition rate was established. The influences of deposition temperature (T), deposition time (t), inlet BCl₃/CH₄ gas ratio (δ), and inlet H₂/CH₄ gas ratio (θ) on deposition rate and microstructure of the coatings were investigated. The optimized deposition parameters were obtained theoretically. The morphologies, phases, microstructure and composition of deposits were characterized using scanning electron microscopy (SEM), X-ray diffraction (XRD), Raman micro-spectroscopy, transmission electron microscopy (TEM), energy dispersive spectra (EDS), and Auger electron spectra (AES), the results showed that different boron carbides were produced by three kinds of deposition mechanisms.