Jeremy Tantrum

RESEARCH INTERESTS

My primary research interests are in statistical learning. I am particularly interested in all clustering methods, especially as applied to the data mining paradigm. I am also interested in modeling social networks, text mining and credit scoring, and in particular interactions between these.

Employment Experience

2004-2005

Senior Statistician at Intelligent Results Inc where my responsibilities included developing models and modeling algorithms for financial scoring.

2003-2004

Post Doctoral researcher in the CSDE developing statistical models for social networks.

Supervised by Mark Handcock

Education

2003

Ph.D. in Statistics, University of Washington, Seattle

Dissertation: Model Based and Hybrid Clustering of Large Datasets

Advisors Werner Stuetzle and Alejandro Murua

1999

M.S., Statistics, University of Washington, Seattle

1997

B.S., with First Class Honors, University of Auckland, New Zealand

Majored in Statistics, also with additional focus in Computer Science and Mathematics

Honors Project: Combining Databases

Advisor Robert Gentleman

Awards and Grants

• CSSS seed grant for research into clustering social networks, 2004.
• IBM Research Student Scholarship, 2003, for Paper submitted to KDD03.
• AAAI Research Student Scholarship, 2002, for Paper submitted to KDD02.
• Senior Prize in Statistics, 1996, University of Auckland.

RESEARCH EXPERIENCE

2003-2004

Post Doctoral Researcher. Modeling Social Networks.

Advised by Professors Mark Handcock and Martina Morris.

I am currently developing models for effectively modeling the structure of social networks. I have worked on Bayesian priors for avoiding degeneracy issues in the models and also a Bayesian mixture model for representing the inherent clustering within networks.

1999-2003

Research Assistant. Clustering of Larger Data Sets.

Advised by Professors Werner Stuetzle and Alejandro Murua.

My dissertation research was on the model based and non-parametric clustering of large datasets. I developed a method for hierarchical fitting a Gaussian mixture model to datasets too large to fit in memory. I also developed a method and theory for bridging the gap between model based clustering and non-parametric clustering. My motivating dataset is a document collection of sixteen thousand news stories.

1999-2000

Research Student. Spatial Modeling.

Advised by Professor Julian Besag.

Developed methods for fitting maximum likelihood estimates to spatially correlated logistic regressions using Monte Carlo Markov Chains. I also looked at the application of this to real agricultural data.

1996

Honors Research Project. Combining Databases.

Advised by Professor Robert Gentleman.

Researched the combining of databases with no records in common, by way of matching on demographic covariates.

Journal Articles

Model-Based Clustering for Social Networks.

J. M. Tantrum, M. S. Handcock and A. E. Raftery. In submission.

Hierarchical Model-Based Clustering of Large Datasets Through Fractionation and Refractionation.

J. M. Tantrum, A. Murua and W. Stuetzle. Information Systems Volume 29, Issue 4 , June 2004, Pages 315-326

Likelihood analysis of binary data in space and time.

J. E. Besag and J. M. Tantrum. In Highly Structured Stochastic Systems. Peter J Green, Nils Lid Hjort and Sylvia Richardson editors, Oxford University Press, April 2003.

Refereed Conference Publications

Assessment and Pruning of Hierarchical Model Based Clustering.

J. Tantrum, A. Murua and W. Stuetzle. 9th Int. Conf. on Knowledge Discovery and Data Mining (KDD03)

Hierarchical Model-Based Clustering of Large Datasets Through Fractionation and Refractionation.

J. M. Tantrum, A. Murua and W. Stuetzle. Proc. 8th Int. Conf. on Knowledge Discovery and Data Mining (KDD02) 2002. 183-190.

Invited Talks

2003

Boeing Research and Technology, Seattle, WA

2001

Classification Society of North America, St Louis, MO

Computing Experience

Languages:

C, HTML, LaTeX, Perl, R, S-PLUS.

Operating Systems:

UNIX, Windows.

Activities

• Managed web site for the Center for Statistics and the Social Sciences (1999 - 2003).
• Served as the Graduate Student Computing Representative (1999 - 2002).
• Provided departmental student software support, University of Washington (2000 - 2003).
• Assisted in computer laboratory, University of Auckland (1995 - 1997).
• Helped lead tutoring of inner city children at True Vine Community Church (2000 - 2003).
• Led Graduate InterVarsity Christian Fellowship bible studies (1999 - 2002).
• Taught introductory statistics course, University of Washington (1997 - 1998).
• Tutored introductory statistics course, University of Auckland (1996).