论文部分内容阅读
提出一种保留模式特征的映射算法,并具体应用于高维的、分量间存在复共线性的复杂化学模式特征的映射,取得良好效果。算法首先提取能最大限度表达复杂化学模式特征的特征变量,并进而采用自组织映射算法将由特征变量组成的特征模式映射到低维空间,使自组织映射结果平面能直观地显示复杂化学模式的特征。算法在具体应用过程中根据复杂化学模式样本的特性,提出了主成分特征映射和分类相关成分特征映射算法。
A mapping algorithm of preserving mode features was proposed and applied to the mapping of complex chemical modes with complex collinearity in high dimension and achieved good results. The algorithm firstly extracts the characteristic variables that can express the features of the complex chemical model to the maximum, and then uses the self-organizing mapping algorithm to map the feature model composed of the characteristic variables into the low-dimensional space so that the self-organizing mapping result plane can intuitively display the features of the complex chemical model . Based on the characteristics of complex chemical model samples, the algorithm proposed in this paper is based on principal component feature mapping and classification related feature feature mapping algorithm.