Comparison of Principal Component Analysis and Linear Discriminant Analysis applied to classification of excitation-emission matrices of the selected biological material

Abstract

Quality of two linear methods (PCA and LDA) applied to reduce dimensionality of feature analysis is compared and efficiency of their algorithms in classification of the selected biological materials according to their excitation-emission fluorescence matrices is examined. It has been found that LDA method reduces the dimensions (or a number of significant variables) more effectively than PCA method. A relatively good discrimination within the examined biological material has been obtained with the use of LDA algorithm.[b]Keywords[/b]: Feature Analysis, Fluorescence Spectroscopy, Biological Material Classification

Keywords

• feature analysis
• fluorescence spectroscopy
• Biological Material Classification