How to get the exact samples inside SOM neurons
8 views (last 30 days)
Show older comments
I have abalysed my data using neural clustering for Self Otganising Maps from the Neural Net Clustering from Machine Learning and Deep Learning app in MATLAB. I am not sure how to know the exact samples inside the neurons after clustering as shown on the sample hits map. For example, I have 7 samples inside one neuron, how do I identify the samples from my original data.The sampe hits are as per attached ![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1359938/image.png)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/1359938/image.png)
0 Comments
Accepted Answer
Priyank Pandey
on 26 Apr 2023
Hi,
To identify the samples from your original data set that are assigned to a particular neuron in your self-organizing map (SOM), you can use the 'vec2ind' function in MATLAB.
The 'vec2ind' function takes as input a matrix of neuron outputs, where each row corresponds to the output of a neuron for a particular input sample. The output of the function is a vector of indices that indicates which neuron had the highest output for each input sample.
To use 'vec2ind' to identify the samples associated with a particular neuron, you need to find the index of the neuron you are interested in on your SOM's hit map, and then find the corresponding row in the output matrix that represents that neuron's output for each input sample. The indices of the input samples that are associated with that neuron can then be obtained from the output of 'vec2ind'.
To know more about how to use vec2ind function refer to the following code:
I hope this helps.
Regards
Priyank
0 Comments
More Answers (0)
See Also
Categories
Find more on Function Approximation, Clustering, and Control 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!