Adaptive noise cancelation by implementing a filter using the ADALINE NN and train it by using the Widrow-Hoff Learning Rule using the noise source x(n), that has corrupted the speech.
Ideally, the neural network is capable of estimating the noise so that e(n) is the reconstructed speech signal.
Classification Results after 4 iterations with learning rate of 0.01
