Classification of Readily Biodegradable Molecules Using Principal Component Analysis and Artificial Neural Network

Abstract

This paper proposes a classification method in order to discriminate readily and not readily biodegradable molecules by means of principal component analysis and Artificial Neural Networks. The data used is taken from UCI machine learning repository. Each sample contains 41 attributes which is too large and will take more time for processing. So principal component analysis is used so that the attributes contributing more information for pattern classification should be given priority over the less contributing attribute. Principal component analysis uses some steps to choose the important attributes. For classification of readily biodegradable data artificial neural network with feed forward back-propagation network is used. The proposed method is effective and reliable with accuracy of 99%.