Ultra high frequency sigmoid and cosine artificial higher order neural networks

This paper developed a new open box and nonlinear model of Ultra High Frequency Sigmoid and Cosine Artificial Higher Order Neural Network (UGC-HONN). This paper also developed a new learning algorithm for UGC-HONN and an UGC-HONN Simulator. Test results show that that UGC-HONN models are 2.8546% to 3.4185% better than Polynomial Higher Order Neural Network models, trigonometric Higher Order Neural Network models, and Sigmoid Polynomial Higher Order Neural Network models.