while using som how to cluster the data into three clusters?

3 views (last 30 days)
I'm using som for clustering in my project. I want the image to be clustered into three clusters. so in this case how to set the values for 'd1' and 'd2' in the function 'selforgmap' ?

Answers (2)

Greg Heath
Greg Heath on 9 Oct 2012
Kohonen's SOM algorithm is for UNSUPERVISED clustering. For supervised clustering he recommends his LVQ and LVQ2 algorithms
Nevertheless, you could try
1. Set d1 or d2 to be 3 and vary the other
or
2. Try initsompc
>> lookfor som
>> lookfor LVQ
Hope this helps.
Greg

Junaid
Junaid on 8 Oct 2012
Try using VLFEAT matlab implementation for Clustering, it has many algorithms with tutorial and example codes. You can easily see visually the clusters with different K.

Categories

Find more on Statistics and Machine Learning Toolbox in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!