MAKHFI.COM
Fascinating World of Neural Nets  
Neural Network Forums
Home PageHome Page : General Neural Networks Topics : Kohonen Map (SOM)
  You are currently not logged in. You can view the forums, but cannot post messages. | HOME | Log In | Register | Search | Help
Post a Reply on This Topic Post a Reply on This Topic

Author Topic: Kohonen Map (SOM)
yuha Posted: 03-May-11 09:53
  Edit Edit
 
Email the Author Mail   View Author's Profile Profile  
I am trying to train a kohonen network to get the optimal weight values. I have input vectors of length 5, and 3 clusters. Can I use this net for medical diagnosis environment. In case yes, the initial weight values are set to random values, my question, how can I deal with the input (symptoms), what kind of values should i give them.

thanking you in advance
 
pejman Posted: 03-May-11 16:05
Delete Delete    Edit Edit
 
Email the Author Mail   View Author's Profile Profile  
It seems you need a classifier (if clustering is already done).
In that case why not using MLP?

As for inputs (symptoms), send me more details.

Cheers,
-Pejman
 
yuha Posted: 04-May-11 02:37
Delete Delete    Edit Edit
 
Email the Author Mail   View Author's Profile Profile  
I am using a type of recurrent neural network for data classification, but this kind of net is a not learning net and needs the optimal weight values to be given by the user in order to provide an accurate results. For that reason I need a learning technique that I can use to get the best weight values (almost the optimal).

My application is the medical diagnosis.
My input vector is a set of symptoms
The output should be a disease according to the given input.

thanking your kind efforts
 
yuha Posted: 04-May-11 02:46
Delete Delete    Edit Edit
 
Email the Author Mail   View Author's Profile Profile  
cont.

I prefer to use a competitive learning mechanism to get the optimal weight values, my question is: Is it possible to use any competitive learning to train my environment which is a medical diagnosis. In case yes, I am giving at initial stage a random weight values, how could I present my input to such competitive learning.
 

Post a Reply on This Topic Post a Reply on This Topic
 

Copyright © 2001-2003 Pejman Makhfi. All rights Reserved.