关键词: 线性判别分析, 类间离散度矩阵, 类内离散度矩阵, 秩限制问题

Abstract: Improved Linear Discriminant Analysis（ILDA） is presented to improve the classification performance of LDA by modifying the original Fisher discriminant criterion，which aims at not only overcoming the rank limitaiton of LDA，but also solving the small size sample problem.This paper mainly discusses the effectiveness of ILDA in overcoming rank limitation.Experiments on some publicly available datasets and one artificial dataset show that the proposed algorithm ILDA has good performance and arrives at the target of getting much more discriminant features.

Key words: Linear Discriminant Analysis（LDA）, between-class scatter matrix, within-class scatter matrix, rank limitation