# A line begins with the character # means comment
# The data file used is m-ibmln99.dat
# I used past three lagged values as input nodes.
#
# load the data into R workspace
x_scan(file='m-ibmln99.dat')
# select the output: r(t)
y_x[4:864]
# obtain the input variables: r(t-1), r(t-2), and r(t-3)
ibm.x_cbind(x[3:863],x[2:862],x[1:861])
# build a 3-2-1 network with skip layer connections and linear output.
ibm.nn_nnet(ibm.x,y,size=2,linout=T,skip=T) 
# print the summary results of the network
summary(ibm.nn)
# compute, then print, the residual sum of squares.
sse_sum((y-predict(ibm.nn,ibm.x))^2)
print(sse)
# setup the input variables in the forecsting subsample
ibm.p_cbind(x[864:887],x[863:886],x[862:885])
# compute the forecasts
yh_predict(ibm.nn,ibm.p)
# select the observed returns in the forecasting subsample
yo_x[865:888]
# compute, then print, the sum of squares of forecast errors
ssfe_sum((yo-yh)^2)
print(ssfe)