Title :
Classification of Listed Companies Based on Kernel Principal Component Analysis and Improved Grid-based Algorithm
Author :
Ren, Hui-xian ; Zhu, Mei-lin
Author_Institution :
Coll. of Manage. Sci. & Eng., Nanjing Univ., Nanjing
Abstract :
After the full understanding of principal component analysis and kernel principal component analysis, the validity of kernel principal component analysis method is proved, and a new grid-based parameter optimization algorithm is introduced. Combining with BPNN, the classification experimental results show that this parameter optimization algorithm can get good results, The performance is superior to traditional methods.
Keywords :
backpropagation; neural nets; optimisation; pattern classification; principal component analysis; stock markets; BPNN; grid-based algorithm; grid-based parameter optimization algorithm; kernel principal component analysis; listed company; pattern classification; stock market; Companies; Data security; Educational institutions; Engineering management; Feature extraction; Investments; Kernel; Optimization methods; Principal component analysis; Stock markets; KPCA; Kernel Function; PCA; Parameter Optimization;
Conference_Titel :
Natural Computation, 2008. ICNC '08. Fourth International Conference on
Conference_Location :
Jinan
Print_ISBN :
978-0-7695-3304-9
DOI :
10.1109/ICNC.2008.544
