Dear all,
dear jcl,

I am designing a strategy which uses Deep Learning with a great number of variables and long sliding window.
I have a sliding window of 128 and 163 features.
This means that I would pass to adviseLong an array of 20864 var.
I can't do it, it crashes.
Do you have any other good idea?
I am thinking to pass only the 163 features and then build the sliding window array from R, but the problem is the prediction part.
How to give to prediction the deepness of the sliding window?
Can you please give me some hint?

Thanks.
Best regards.