Contact Information:
 Operations Research and Financial Engineering 
 Princeton University
 kmtan @ princeton.edu
I am a postdoctoral research associate supervised by Professors Han Liu from Princeton University and Tong Zhang from Rutgers University. Prior to this, I obtained my PhD degree in the Department of Biostatistics at University of Washington, under the supervision of Dr. Daniela Witten.

I am interested in developing statistical machine learning methods for analyzing high-dimensional data sets arising from genomics, neuroscience, and other areas of biology.  More specifically, I develop unsupervised learning methods such as probabilistic graphical models, cluster analysis, and dimension reduction to uncover patterns from massive data set.    

I am currently collaborating with neuroscientists at the Princeton Neuroscience Institute to study the reconfiguration of brain connectivity network under natural continuous stimulus.  See more details here.

Featured Papers

Sparse Generalized Eigenvalue Problem: Optimal Statistical Rates via Truncated Rayleigh Flow
Tan KM, Wang Z, Liu H and Zhang T (2016).
Under Revision in JRSSB.  

Replicates in High Dimensions, With Application to Latent Variable Graphical Models
Tan KM, Ning Y, Witten D and Liu H. (2016).
Biometrika, 103(4):761-777  
​
Statistical Properties of Convex Clustering
Tan KM and Witten D. (2015). Electronic Journal of Statistics 9:2324-2347

Learning Graphical Models with Hubs   
Tan KM, London P, Mohan K, Lee S-I, Fazel M and Witten D. (2014). Journal of Machine Learning Research 15(Oct):3297-3331
​
Sparse Biclustering of Transposable Data
Tan KM and Witten D. (2014).  Journal of Computational and Graphical Statistics (23)4:985-1008

R Packages

hglasso (CRAN link)
An R software package for estimating graphs with hub nodes
 
sparseBC (CRAN link)
An R software package for performing sparse biclustering 
​
latentGraph (coming soon)
An R software package for estimating latent variable graphical models
​

Presentations

Poster on Dynamic Reconfiguration of Brain Network under Natural Continuous Stimulus
Slides on Replicates in High Dimensions, With Application to Latent Variable Graphical Models

Slides on Statistical Properties of Convex Clustering
Poster on Learning Graphical Models with Hubs