I am a bit confused about how can i visualize clusters in a Self Organizing Map. The input data is a set of images, where each image is an english alphabet in some font. Now if i have to visualize the final organization and clusters, where each alphabet/letter that is similar is shown as a cluster, how can i achieve this?

I previously thought that SOM algorithm does it for you and my current visualization shows the image generated using the weight vector after each iteration for every single output node.

I realized my approach to visulize doesn't show clusters but it shows how each output node is converging towards some input image. Also one or more output nodes may be converging towards different input images.

I have read about U-matrix approach but still unclear if i will be able to visualize similar inputs as seperate clusters.

My question is that if i have a data set of letters where each letter belongs to the set X={A,B,C,D,E,F} and each letter is in 6 different fonts, hence forming a data set of total 36 letters.

How can i visualize the 6 clusters where each cluster has 6 letters in it ? Also does the dimension of the lattice be equal to that of the inputs in the dataset for better visualization ?

Any help will be greatly appreciated. Thanks in advance


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.