Neural network model selection for financial time series prediction (Q1861629)

This is a case study devoted to feed-forward neural network selection for the prediction of the monthly mortgage loans volume (MV) purchased in the Netherlands. The predictors are the inflation rate, interest rate, number of dwellings for sale in the housing market, housing price and GDP. The authors use a network with one hidden layer, logistic link units activation function and linear output units. The S-PLUS neural network library of Venables and Ripley was used.