How can we understand spectral clustering algorithm in a more generalized way? The research papers mention it in more mathematical form. What could be a layman's approach to understand it by overlooking mathematical equations?
A simple answer is that, spectral clustering is useful when features that represent different values/items are present in feature space in linearly mixed form (for example in spheres around one another). In such case, simple clustering will not help and you will need spectral clustering.
Hello Anupam,
Here i have given some Information it may be helpful to you i think so.
Algorithms that cluster points using eigenvectors of matrices derived from the data
Obtain data representation in the low-dimensional space that can be easily clustered
Variety of methods that use the eigenvectors differently
Empirically very successful
Authors disagree:
Which eigenvectors to use
How to derive clusters from these eigenvectors
Two general methods
Partition using only one eigenvector at a time
Use procedure recursively
Example: Image Segmentation
Uses 2nd (smallest) eigenvector to define optimal cut
Recursively generates two clusters with each cut
- Badges
- Science topic