Predicting values with TemporalMLDataSet? #159
Description
Dear Jeff,
first, I want to thank for your good work here.
I posted the following item some month ago in the old forum. I would like to ask, if you could spend some time on checking it:
I'm a relatively new user of the Encog framework, read and worked through "Programming Neural Networks with Encog 3 in Java" completely and searched on the internet resources, but didn't find the answer for my following question:
When I try the example from FAQ 3.01 (http://www.heatonresearch.com/faq/3/1), I would like to know, how I can compute the prediction for the same input data.
The training data uses "TemporalMLDataSet" with two inputs (number of sunspots and deviation).
This is ok and clear to me.
The example doesn't contain a prediction (or training) code part, so I checked the similar example from the above mentioned book, where the "BasicNetwork.compute(MLData)" method is used for prediction. But the MLData can not hold two inputs of a time series, am I right? So this way, I could only input the number of sunspots into to the network, not the deviation.
How can I predict a time series, when I have two (or maybe more) input time series? I saw questions like this many times in the forum, but never figured out an answer, I could use for this task.
Any help would be appreciated.
best regards,
Winfried