ISSN : 2319-7323

INTERNATIONAL JOURNAL OF COMPUTER SCIENCE ENGINEERING

ABSTRACT

Title	:	Enhanced MLP classifier performance using variant of Back Propagation on Modified Cervical Pap smear Data
Authors	:	Dr. K. Usha Rani, Dr. K. Hemalatha
Keywords	:	Neural Network; Multi-Layer Perceptron; Back Propagation algorithms; Cervical Cancer data.
Issue Date	:	Nov-Dec 2019
Abstract	:	Multi-Layer Perceptron (MLP), the most popular Neural Network (NN) widely used in Medical applications. Training process is an important pace which can influence the performance of NN. Learning Algorithm plays a vital role in training NN. Back Propagation (BP) algorithm is commonly used supervised Learning algorithm for training MLP. Several variants of BP algorithms are available in Literature. Selection of suitable BP algorithm is significant to enhance the performance of MLP. In this study widely used BP algorithms: Bayesian Regularization (BR), Resilient back propagation (Rprop), Scaled Conjugate Gradient (SCG) and Levenberg-Marquardt (LM) are considered to train the MLP on Modified Cervical Pap smear (MCPS) Data and the results are analyzed. The best BP variant which enhanced the MLP performance is identified.
Page(s)	:	255-260
ISSN	:	2319-7323
Source	:	Vol. 8, No. 6