I understand that you want to use "narxnet" to predict the values of a time series. The “inputDelays” parameter in a NARX neural network determinesthe number of past time steps of the input time series that the network considers when making predictions for future time steps.
For example, if “inputDelays” is set to 1:2 with inputs “X” and Targets “T”, it means that the model uses the set "{x(i-1), x(i-2)}” to predict "t(i)".
If you need more information about the "inputDelays" parameter, I recommend referring to the following resources: -
I hope this helps!
