Date: Jan 24, 2013 7:25 PM
Author: Greg Heath
Subject: Re: DTDNN
"saba " <firstname.lastname@example.org> wrote in message <email@example.com>...
> Hi Greg,
> I have a question regarding the use of distributed time delay neural network.I wondered whether the input and target should be the same set of data or not for training?? I mean if I have different data sets of wind, solar radiation and temperature, and I want to predict one step ahead of temperature using past data (100 data). Can I train the network with data input wind and output temperature?
You can use any inputs that are assumed to be correlated to the output.
If the inputs and outputs are not measured within the same time interval, how can you make that assumption?
Hope this helps.