home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai.neural-nets
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!spool.mu.edu!news.nd.edu!mentor.cc.purdue.edu!noose.ecn.purdue.edu!cornsilk.ecn.purdue.edu!shashem
- From: shashem@cornsilk.ecn.purdue.edu (Sherif Hashem)
- Subject: Re: NN analysis
- Message-ID: <1992Nov16.031030.2415@noose.ecn.purdue.edu>
- Sender: news@noose.ecn.purdue.edu (USENET news)
- Organization: Purdue University Engineering Computer Network
- References: <1dvfvkINN2tv@manuel.anu.edu.au>
- Date: Mon, 16 Nov 1992 03:10:30 GMT
- Lines: 37
-
- In article <1dvfvkINN2tv@manuel.anu.edu.au> shuping@andosl.anu.edu.au (Shuping RAN) writes:
- >Hello,
- >
- > Currently I am trying to understand how a trained NN performed
- >the given task. What is its internal funtionality, whether its internal
- >parameters relate to the real parameters of the given problem in some
- >way.
- >
- > Could some one give me some ideas of how to analyse a trained NN,
- >or some references.
- >Thank you in advance,
- >
- >-Shuping RAN
-
- In the area of using NNs for function approximation, C. Klimasauskas
- wrote an article in Dr. Dobb's Journal, April 1991, 16-24, by
- the title: Neural Nets Tell Why.
-
- Also I used sensitivity analysis to study the input-output relationships
- learnt by a trained network. I published part of my work in a paper titled:
- "Sensitivity Analysis for Feedforward Artificial Neural Networks
- with Differentiable Activation Functions," in the proceedings of the`
- 1992 International Joint Conference on Neural Networks in Baltimore,
- IEEE, NY, I:419-424.
-
- Sherif Hashem
-
- School of Industrial Engineering
- 1287 Grissom Hall, Purdue University
- West Lafayette, IN 47907-1287.
- shashem@ecn.purdue.edu
- (317)494-0440 (O), (317)494-1299 (Fax).
- --
- Sherif Hashem
-
- School of Industrial Engineering
- 1287 Grissom Hall, Purdue University
-