/Spectral-Clustering

Spectral Clustering is a technique to cluster data which finds use in community detetction applications

Primary LanguageJupyter Notebook

Spectral-Clustering

Spectral Clustering is a technique to cluster data which finds application in community detection applications. It is equivalent to finding a min-cut of a graph where edge (or their weights) represents closeness which is subjective to data and context.

Dolphins Dataset - notebook_spectral_clustering_1.ipynb :

Dolphins Community Detetction, Two Circles Problem (Spectral clustering from scratch)

Unclustered

alt text

Clustered

alt text

data: data.zip

Identify Player Team based on Interaction - Barca_madrid.ipynb (everything form scratch) :

Barca - Real Madrid clustering problem (everything from scratch - eigenvalues,eigenvecs, spectral clusetring)

Unclustered

alt text

Clustered

alt text

data: lclassico.gml