USP Electronic Research Repository

A hybrid approach for modeling high dimensional medical data

Sharma, Alokanand and Onwubolu, Godfrey C. (2007) A hybrid approach for modeling high dimensional medical data. [Conference Proceedings]

Full text not available from this repository.

Abstract

This work presents the application of hybrid PCA and LDA to modeling high dimensional medical data, which is a real-life problem. For modeling and classifying medical data, we adopted this combination of two stage PCA and LDA procedure which is also known as Fisherface technique. During the training phase we applied this combination for extracting features from medical data. In the classification stage we introduced weighting ratio which is used with the conventional Euclidean distance measure to classify a given sample. For brevity we call this technique the weighted distance Fisherface technique. The presented technique shows promising results for medical data when compared with standard GMDH technique; in the two problems taken from the machining learning databases, the presented approach performed better than the standard GMDH.

Item Type: Conference Proceedings
Subjects: T Technology > TJ Mechanical engineering and machinery
Divisions: Faculty of Science, Technology and Environment (FSTE) > School of Engineering and Physics
Depositing User: Ms Mereoni Camailakeba
Date Deposited: 26 Apr 2007 14:23
Last Modified: 23 Jul 2012 19:15
URI: http://repository.usp.ac.fj/id/eprint/4260
UNSPECIFIED

Actions (login required)

View Item