KMID : 1059519920360010113
|
|
Journal of the Korean Chemical Society 1992 Volume.36 No. 1 p.113 ~ p.124
|
|
Classification of Korean Ancient Glass Pieces by Pattern Recognition Method
|
|
Lee Chul
Case Myung-Zoon Kim Seung-Won Kang Hyung-Tae Lee Jong-Du
|
|
Abstract
|
|
|
The pattern recognition methods of chemometrics have been applied to multivariate data, for which ninety four Korean ancient glass pieces have been determined for 12 elements by neutron activation analysis. For the purpose, principal component analysis and non-linear mapping have been used as the unsupervised learning methods. As the result, the glass samples have been classified into 6 classes. The SIMCA (statistical isolinear multiple component analysis), adopted as a supervised learning method, has been applied to the 6 training set and the test set. The results of the 6 training set were in accord with the results by principal component analysis and non-linear mapping. For test set, 17 of 33 samples were each allocated to one of the 6 training set.
|
|
KEYWORD
|
|
|
|
FullTexts / Linksout information
|
|
|
|
Listed journal information
|
|
|