Train a neural network with a single hidden layer and 10 hidden units to data: 100 samples generated from Y = o(afX) + (ax) +0.3Z, where o is the sigmoid function, Z is standard normal, XT = [X, X]T each X, being independent standard normal, and a = [3,3]T, a2 = [3, 3]. Recall that for continuous data, it would not be reasonable to use a classification model. Neural networks can be used for regression problems just like they can be used for classification, ensure you have selected software which can support this case. Generate a test sample size 1000, and plot the training and test error (MSE) curves as a function of the number of training epochs (recall, an epoch is an iteration over the entire dataset) for different values of the weighted decay parameter (some packages call this the 12 regularization rate). Discuss the overfitting behavior in each case. Now vary the number of hidden units in the network from 1 up to 20, and determine the minimum number needed to perform well for this task. Activate Windows Go to Settings to activate Wind