Neural network

Problem: Design a multilayer perceptron for function approximation y=f(x) based on a given training set. Network structure: 2 input neurons: x and 1 (bias), one hidden layer of m neurons, one output neuron. Activation function of hidden neurons : g (h )  tanh( h )  Activation function of output neuron: linear. exp( h )  exp(  h ) exp( h )  exp(  h ) Data sets pairs {x,y}: - approx_<x>[login to view URL], - training set: 100 examples {x,y}uniformly distributed in the interval (0, 15) (first row – x coords, second row – y values) - approx_<x>_test.asc. - test set: 400 examples {x,y} uniformly distributed in the interval (0, 15).