Hi Guys,
I have run the Neural network for the purpose of classification.
however, after training the data , I have found the mis-classification
numbers are quit high.
Just wondering what do your guys deal with this when this thing
happened?
What factors would influence the results.
I am using the results from text mining to train my data. so I used
the SVDs as input to train data.
so my input is 100 SVDs and target is the classification categories it
is about 23 clusters. i have tried setting the network from 6 hidden
neuro to 40 neuros.
please help.
any suggestions and ideas are fine.
thanks.