基于余弦角距离的主成分分析与核主成分分析

PCA和KPCA都是基于欧氏距离提出的，这种距离对离群数据点比较敏感，而余弦角距离对离群数据更为鲁棒，在很多情况下具有更好的性能。充分利用余弦角距离的优势，提出两种新的特征抽取算法——基于余弦角距离的主成分分析（PCAC）和基于余弦角距离的核主成分分析（KPCAC）。在YALE人脸数据库与PolyU掌纹数据库上的实验表明，PCAC比PCA取得了更好的效果，KPCAC也表现出了很好的性能。

Principal component analysis and kernel principal component analysis based on cosine angle distance

Principal Component Analysis（PCA） and Kernel Principal Component Analysis（KPCA） are both proposed based on Euclidean distance which is sensitive to outlier.Cosine angle distance is more robust to outlier and has better performance in many cases.This paper utilizes the superiority of cosine angle distance sufficiently and proposes two new feature extraction algorithms——Principal Component Analysis based on Cosine（PCAC） angle distance and Kernel Principal Component Analysis based on Cosine（KPCAC） angle distance.Experiments on YALE face database and the PolyU palmprint database show the superiority of PCAC over PCA and the effectiveness of KPCAC.