Is there a maximum number of delays?

Hi, May i know whether there is any maximum number of delays i can have for neuron networks time series tools?

Answer 1

[Greg Heath]

Statistically significant nonnegative input delays can be obtained using the input/target cross-correlation function.

Statistically significant positive feedback delays can be obtained using the target autocorrelation function.

The optimal number should be a subset of those found above.

I have posted tens of examples. Search using

 greg nncorr

Hope this helps

Thank you for formally accepting my answer

Greg

[WT]

I don't quite understand what you mean..Do you mind sending me the links of your examples?

[Greg Heath]

See the white box at the top of the page that says

 Search MATLAB Answers?

Type in

 greg nncorr

Then click on the square button on the right.

Answer 2

[Star Strider]

To the best of my knowledge, no, although it may be practically limited by your data set charactersitics.

Nothing in the documentation that I can find limits them.

[WT]

So i assume there is no optimal numbers of delays as well?

[Star Strider]

That likely depends on your dataset characteristics. The default is 2, but you may want more. It depends on how many of your data you want to consider. You will have to experiment to find the number that is small enough to be efficient and large enough to be accurate.

[WT]

But the values of MSE keep changing so i have difficulty tracking the optimal numbers of delays. so there isn't any formula where i can estimate the value of delays?