M. Zhang (USA)
Neural Networks, Higher Order, six/x, SXSPHONN, Simulator, Rainfall Estimation *This research is supported by Christopher Newport University (CNU), The College of Liberal Arts and Sciences, Dean’s Office Grant for Spring 2005 and CNU Applied Research Centre 2005 research funding.
Traditional methods of data simulation are high inaccurate. Artificial neural networks have strong pattern finding ability. But presently the solution to choice of the optimal neural network model for simulation and prediction is still not found. In this paper a data simulation system based on the sinx/x and SINX Polynomial Higher Order Neural Network (SXSPHONN) has been developed. The SXSPHONN model gives one more chance to find out the optimal neural network model for simulation and prediction. The paper also proved that SXSPHONN models are convergence when simulate XOR data. Rainfall data has been tested in the paper. Based on the test results, SXSPHONN model is 0.999% better than Polynomial Higher Order Neural Network (PHONN) model and 0.967% batter than trigonometric Polynomial Higher Order Neural Network (THONN) model in our rainfall case.
Important Links:
Go Back
