This paper presents a comparative study between two

\r\nneural network models namely General Regression Neural Network

\r\n(GRNN) and Back Propagation Neural Network (BPNN) are used

\r\nto estimate radial overcut produced during Electrical Discharge

\r\nMachining (EDM). Four input parameters have been employed:

\r\ndischarge current (Ip), pulse on time (Ton), Duty fraction (Tau) and

\r\ndischarge voltage (V). Recently, artificial intelligence techniques, as

\r\nit is emerged as an effective tool that could be used to replace

\r\ntime consuming procedures in various scientific or engineering

\r\napplications, explicitly in prediction and estimation of the complex

\r\nand nonlinear process. The both networks are trained, and the

\r\nprediction results are tested with the unseen validation set of the

\r\nexperiment and analysed. It is found that the performance of both the

\r\nnetworks are found to be in good agreement with average percentage

\r\nerror less than 11% and the correlation coefficient obtained for the

\r\nvalidation data set for GRNN and BPNN is more than 91%. However,

\r\nit is much faster to train GRNN network than a BPNN and GRNN is

\r\noften more accurate than BPNN. GRNN requires more memory space

\r\nto store the model, GRNN features fast learning that does not require

\r\nan iterative procedure, and highly parallel structure. GRNN networks

\r\nare slower than multilayer perceptron networks at classifying new

\r\ncases.<\/p>\r\n","references":null,"publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 88, 2014"}