基于QR分解和支持向量的伪逆LDA

摘要

为了降低计算复杂性,本文提出了三种基于QR分解和支持向量的伪逆线性判别分析法.通过对UCI数据库中的Wine和Iris数据进行对比实验,验证了本文所提算法对分类问题的有效性和实用性.%In order to reduce computational complexity,three kinds of pseudo-inverse linear discriminant analysis methods based on QR-decomposition and support vectors are presented in this paper.Based on Wine and Iris datasets taken from UCI database,experiments compared with pseudo-inverse LDA show that the effectiveness and practicality of the three methods for classification results.