Getting Target Data from Given Input Matrix Neural Networks

I am given an input data matrix representing emails. Each row corresponds to one email and the columns correspond to attributes of that email as numeral values. The last column is my target column, which classifies whether or not the E-Mail is a spam using 0 (not spam) and 1 (spam).

I now want to create and train a neural network to identify whether an email is spam or not based on this last column using the Neural Networks Toolbox. How to I create the target data from my file?

I am pretty confused about how to create files in matlab in general, so a detailed description would be appreciated.

Thanks!

Answer 1

[Greg Heath]

N pairs of I-dimensional "I"nput vectors and corresponding O-dimensional "O"utput target vectors stored in matrices input and target:

 [ I N [ = size(input) 
 [ O N ] = size(target)

 output = net(input);
 [ O N ] = size(output)

Hope this helps

Thank you for formally accepting my answer

Greg

[Karo Gaebe]

so I would have N=4061 (this is the amount of emails I have data for), I=57 and O=1? what do the command lines above do? I also still do not understand how this helps me to extract the data from my matrix file

[Greg Heath]

That is not a complete code.

The first two explicitly tell you the size of your data matrices so that you can verify that you have not accidently transposed a matrix or otherwise made a mistake.

The next is a post training command that yields the output y from a TRAINED net.

The last is not typically in codes. I just wanted to emphasize that the output is the same size as the target.

To see a full classification codes that include training, see the documentation examples via the commands

 help patternnet
 doc patternnet

To see more practical classification codes search both the NEWSGROUP and ANSWERS using

 greg patternnet

To obtain more MATLAB data for practice use the commands

 help nndatasets
 doc nndatasets

Hope this helps

Greg