use of LVQ paradigm in neural net

input= analog value (e.g 2) target= 8 bit digital value ( e.g 00000010) i want to use LVQ to train my dataset of 5000 samples. I am getting errors for SPARSE matrix and says unequal vector sizes when I use 'ind2vec' command

Can LVQ be used for the vectors containing binary values, '1' and '0'???

Answer 1

[Greg Heath]

I'm pretty sure it should work.

What distance measure are you using?

Box? Manhattan? Link?

Do all of the vectors the same number of digits?

Hope this helps.

Greg