Development and performance evaluation of improved classifiers of microarray data

High dimensional attributes associated with Cancer patients pose problem in accurate classification of types of cancers. For achieving this objective different feature reduction schemes with a set of simple classifiers have been suggested in this paper. Firstly the dimension have been reduced using Principal Component Analysis(PCA) and proposed to be Discrete Cosine Transform(DCT).Then the reduced dimensions are used to design intelligent classifiers using different Functional Link Artificial Neural Network (FLANN). The performance of proposed model has been evaluated using real life data, using PCA or DCT methods and using three FLANN classifiers. The simulation results demonstrate that the reduction based chebyshev classifier perform the best compared to other alternatives.