Automatic classification of A. paeoniifolius species from DNA fingerprints of Amorphophalus Genus

This work aims at designing a classification system for automatic identification of A paeoniifolius species from the DNA fingerprints of Amorphophalus Genus. Four species, namely A. krausei, AmacrorhizusCraib, A. muelleri and A. paeoniifolius are considered for the experiment. Reduction of the dimensionality of the initial data matrix is done using Principal Component Analysis (PCA). During training stage, Artificial Neural Network based classifiers with back propagation learning algorithm, are designed with 5-fold cross validation experiment. The classification performance of the trained networks, that a DNA fingerprint of an unknown species is A. paeoniifoilius or not, are finally evaluated over an independent validation set.