An Modified Error Function for the Complex-value Backpropagation Neural Networks

The complex-valued backpropagation algorithm has been widely used in fields dealing with telecommunications, speech recognition, and image processing with Fourier transformation. However, the local minima problem usually occurs in the process of learning. To solve this problem and to speed up the learning process, we propose a modified error function. we added a term to the conventional error function, which is connected to the hidden layer error. It can harmonize the update of the weights connected to the hidden layer and output layer. We have applied this method to the detection of symmetry problem and a real classification task. The simulation results show that the proposed algorithm is capable of preventing the learning from sticking into the local minima and of speeding up the learning. Keywords—Modified error function, Complex-BP, learning, local minima