Title :
Factor Analysis Algorithm with Mercer Kernel
Author :
Xia Guo-en ; Shao Pei-ji
Author_Institution :
Univ. of Electron. Sci. & Technol., Chengdu
Abstract :
Nonlinear factor analysis method was studied by Mercer kernel function which can map the data in the original space to a high-dimensional feature space, and a comparison with the related method kernel principle component analysis (KPCA) was made. It is pointed that the best error rate in handwritten digit recognition by kernel factor analysis (KFA) with varimax (4.2%) is competitive with KPCA (4.4%). The results indicate that KFA with varimax could more accurately image handwritten digit recognition and could be an effective measure for studying pattern recognition.
Keywords :
handwriting recognition; image recognition; principal component analysis; support vector machines; Mercer kernel function; high-dimensional feature space; image handwritten digit recognition; kernel principle component analysis; nonlinear kernel factor analysis algorithm; support vector machine; Algorithm design and analysis; Functional analysis; Handwriting recognition; Image recognition; Information analysis; Kernel; Pattern recognition; Performance analysis; Space stations; Space technology; Kernel factor analysis(KFA); Kernel principal component analysis(KPCA); Support vector machine (SVM);
Conference_Titel :
Intelligent Information Technology and Security Informatics, 2009. IITSI '09. Second International Symposium on
Conference_Location :
Moscow
Print_ISBN :
978-1-4244-3580-7
DOI :
10.1109/IITSI.2009.55
