DocumentCode :
2666785
Title :
A pattern recognition method using projection pursuit
Author :
Liu, B. ; Shen, Zhe ; Sun, Zhongyuan
Author_Institution :
Dept. of Electron. Technol., Nat. Univ., of Defense Technol., Hunan
fYear :
1990
fDate :
21-25 May 1990
Firstpage :
300
Abstract :
The authors discuss a method called projection pursuit, which can compress dimensions of feature space. They define a projection index which is suitable for practical use and derive a way to obtain the optimum projecting matrix, i.e. one that maximizes the projection index. The results obtained can be used to project multifeature patterns can be projected to one-dimensional or two-dimensional space, and they will be more easily recognized
Keywords :
computerised pattern recognition; computerised picture processing; digital simulation; military computing; feature space dimensions compression; multifeature patterns; one-dimensional space; pattern recognition; projection pursuit; simulation; two-dimensional space; Data analysis; Pattern recognition; Scattering; Statistical analysis; Sun;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Aerospace and Electronics Conference, 1990. NAECON 1990., Proceedings of the IEEE 1990 National
Conference_Location :
Dayton, OH
Type :
conf
DOI :
10.1109/NAECON.1990.112785
Filename :
112785
Link To Document :
بازگشت