A Study on Design of Discriminant Classification System for the Automatic Classification of Documents

문헌의 자동분류를 위한 판별분류 시스템 설계

This study suggests two hypotheses and verifies them. First hypothesis is that discriminant analysis which is a statistical technique can be used to classify documents on the subject of organic chemistry by nine subareas. Second hypothesis is that discriminant analysis is superior to cluster analysis in classifing objects by fixed categories.