Comments on: Java Neuroph tutorial – The code classifier

By: Guil

Guil — Wed, 28 Sep 2011 15:11:15 +0000

Hi,
This project sound great!
I was wondering though how you get the percentage of certainty as the output is a discrete value isn’t it?

By: Jack

Jack — Sun, 21 Aug 2011 16:49:27 +0000

Any chance we could see the source code of how you did this? Would love to try using this code!

By: CertPal

CertPal — Mon, 27 Jun 2011 05:40:16 +0000

@Qkyrie It is a standard multilayer perceptron. The number of inputs vary based on the number of unique keywords

By: Qkyrie

Qkyrie — Sat, 18 Jun 2011 15:10:44 +0000

I’m very interested in this project, would it be possible to give some more info on what type of network you used (layers etc), as I’m trying to solve a relatively equal problem (classification of hash representations of words).

I would appreciate a response as fast as possible

kind regards
Q

By: CertPal

CertPal — Tue, 27 Jul 2010 16:27:34 +0000

@brainless

Yeah the possibility of wrongly classifying input data is great when the input is not pre-formatted. For example the input string “The code classifier has been release to the public” will result in the classification – java

Thanks for trying out the tool though

By: brainless

brainless — Tue, 27 Jul 2010 12:40:03 +0000

in fact it’s 82%

By: brainless

brainless — Tue, 27 Jul 2010 12:39:25 +0000

x86 assembly (wikipedia hello world sample) looks like assembly at 56% :p