A Radial Basis Function Artificial Neural Network Test for Neglected Nonlinearity

We propose a test for neglected nonlinearity that uses an artificial neural network. We use radial basis functions for the `hidden layer' with basis function centers and radii chosen from the sample data set and selected on the basis of information criteria. The procedure is straightforward to implement and out-performs the random network test proposed by Lee, White, and Granger (1993).