一种有效的加权PCA物质分类方法

An efficient weighted PCA method in material classification

利用经典PCA算法对数据进行分类,所得到的前两个主成分不能代表原始数据的大部分信息,使得分类效果不明显。为了克服这一缺点,提出了一种采用数据预处理技术的加权PCA分类方法,达到较明显的分类效果。该方法首先对样本数据进行归一化处理,将不同尺度的数据规约到同一范围,在此基础上,计算出每个样本在整个数据集合中的权值,接着对归一化后的数据进行均值化或去均值的处理,然后对处理后的样本数据乘以权值来体现其重要程度,最后再利用奇异值分解过程实现主成分分析。计算机仿真结果表明,所提出的方法能很好的将数据分类,与传统PCA法相比,优势明显。

The first two principal components after using classical PCA algorithm to classify the data can not represent the most information of original data so that the effect is not very obvious. In order to overcome the shortcoming, a weighted PCA method adopted the data pre-processing technology in material classification was presented. The method firstly normalized the sample data to make the different scale data to a same range, and then calculated the weight of each sample data in the data set. Based upon that, the sample data which was processed by the technique of equalization or mean removal was multiplied by weight to highlight its importance. Finally, singular value decomposition process was utilized to realize principal component analysis. The computer simulation results show that the proposed method can classify the data well and its advantages are obvious bv eamnarison with traditinn PCA methnd