Substitute input data in Modeler

Here is the explanation of what I am attempting to do with modeler.

We are attempting to use a neural net model to generate outputs
associated with a series of different/select input variable.

We have a series of weather data that need to be substituted into a
basic data set one at a time and executed through the neural net.

We would also like to be able to substitute the data based on a random
draw from a discrete distribution of the weather data and repeat that
operation up to a 1000 time for each line of data in the basic data set.

Below is a screen shot of the basic data set. Desire is to run each
row in the basic data through the neural net, substituting the 10 rows
of Temp, Prec, and Rad data one at a time, and do this for every row in
the basic data set. The output from the neural net would have 100 row
of data (10 row from basic data with each row of data having the 10 row
of data from the substitute data).

We would also like to be able to replace the data based on doing a
random draws against a discrete distribution of the substitute data. In
this case we would like to run at least 1000 iterations of each row in
the basic data through the neural net to generate output distribution.![alt text][1]

[1]: /answers/storage/temp/13242-example.png

Related: