Spectral Clustering
Add to your list(s)
Download to your calendar using vCal
If you have a question about this talk, please contact Colorado Reed.
The term “Spectral Clustering” refers to a family of clustering algorithms related by the fact they exploit methods from spectral graph theory. These methods have proved popular in the machine learning community and are relatively straightforward to implement. This talk will be primarily a tutorial on basic results of spectral graph theory combined with an overview of various ways these techniques can be used for clustering. We will cover the basic properties of graph Laplacians, the relationship between the first eigenvector and the normalized cut, and some simple techniques to obtain clusterings from the eigenvectors of the Laplacian. We’ll then discuss the Nystrom method, an approximate method which is faster and can generalize to points outside the original graph. Since this is a tutorial, no prior reading will be required.
This talk is part of the Machine Learning Reading Group @ CUED series.
This talk is included in these lists:
Note that ex-directory lists are not shown.
|