Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
Self-Oranizing Feature Maps Questions

Subject: Self-Oranizing Feature Maps Questions

From: Ahmad Ammari

Date: 6 Nov, 2008 14:36:02

Message: 1 of 1

I am new to neural networks toolbox and self-organzing feature maps (SOFM) implementation in Matlab. So bare with me please:

I am simply trying to cluster (group) a collection of text documents into distinct groups to study the relationships between them. I have 197 documents. I represented each document with a set of terms (656 terms) and each term with a weight (minimum 0 and maximum 1). In this case I ended up having a term - document matrix consisting of 197 rows (documents) and 656 columns (terms weights).

Now, my questions:

1) What should be the size of my self-organizing map? should I create a map of 197 neurons? with each neuron represents a document? Or there is no relationship between the number of documents I have and the number of the neurons in the map?

2) How many input vectors should be created? And should each input vector contain 656 elements to correspond to the weights of terms in each document?

3) will my data (the matrix I have) sit on the map or it will just be a collection that an input vector is chosen from in each iteration? In the latter case, if my data is not the actual neurons of the map, then how and from where can I bring the initial weights of the neurons of the map?

Thanks very much in advance...

Tags for this Thread

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us