Typically, MATLAB equations are multidimensional.
For regression, classification and time series, all you have to do is create the multidimensional input matrix of size [ I N ] and corresponding multidimensional output target matrix of size [ O N ].
Given input and target, the fewest number of statements, 2, would be
[ net trainingrecord output error ] = train(fitnet,input,target);
Rsq = 1 - mse(error)/mean(var(target',1))
which uses default values that are available via the command
net = net % NO SEMICOLON
The examples for the help and doc commands use a few more statements and can be helpful. UNFORTUNATELY, however, they do not
1. show data size
2. typically, yield a numerical performance value.
3. explicitly initialize the RNG so that numerical results
can be duplicated
Most of the useful auxiliary information will be available in the training record.
Although lists of MATLAB example datasets for regression, clustering, classification, and timeseries can be obtained can be obtained via
help nndatasets
doc nndatasets
sizes are not directly available.
However, sizes are directly available via these posts
http://www.mathworks.com/matlabcentral/newsreader/view_thread/337914
https://www.mathworks.com/matlabcentral/newsreader/view_thread/339984
http://www.mathworks.com/matlabcentral/newsreader/view_thread/338208
and, posted training examples obtained via commands like
greg fitnet
greg patternet
etc
in BOTH the NEWSREADER and ANSWERS.
Hope this helps.
Thank you for formally accepting my answer
Greg
