Return to my homepage

MATH 203: Applied Mathematics, Computing & Statistics Projects (CAMCOS)

Spring 2018, San Jose State University

Acknowledgment

This project, in continuation of Spring 2017, conducts research on scalable spectral clustering. We gratefully acknowledge Verizon Wireless for their generous support.

Toy data

The 20 newsgroups is our main data set.

We will also use the data sets available here: [link].

References

Lecture notes

SVD, dimensionality reduction, and clustering [PDF]

Overview of document clustering

A Survey of Text Clustering Algorithms [Link]

Dimensionality reduction of document data

Latent Semantec Indexing (LSI) [paper]
Locality Preserving Indexing (LPI) [Link]

Spectral clustering

Normalized cuts [paper] [software]
NJW algorithm [paper]
Self-tuning spectral clustering [paper] [webpage]
A tutorial on spectral clustering [paper]
Diffusion maps [lecture] [NIPS paper] [ACHA paper] [PNAS paper]

Landmark based spectral clustering (LSC)

Papers: [AAAI version] [Journal version]
Matlab code: link (download also the three functions bestMap, hungarian, MutualInfo under clustering evaluation)
Websites: [Deng Cai] [Xinlei Chen]