Abstract Details
Activity Number:
|
544
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
Abstract - #310000 |
Title:
|
Branching Out with Level Set Trees: Generalizing Beyond Densities and Enabling Interactive Data Analysis
|
Author(s):
|
Brian P. Kent*+ and Alessandro Rinaldo and Timothy Verstynen
|
Companies:
|
Carnegie Mellon University Department of Statistics and Carnegie Mellon University and Carnegie Mellon University Dept of Psychology and Ctr for the Neural Basis of Cognition
|
Keywords:
|
nonparametric density estimation ;
level set trees ;
clustering ;
functional data analysis
|
Abstract:
|
Level set trees provide a statistically principled framework for visualizing and clustering high-dimensional and complex data. We improve and generalize level set trees in a traditional probability density-based setting by showing how to use them interactively for data analysis and by indexing the tree height by the empirical probability content rather than level values. Next we extend the applicability of level set trees by generalizing the method to data sets that lack bona fide probability densities. In functional data analysis for example, we use pseudo-densities to construct level set trees for infinite-dimensional data and show that this technique produces reliable and accurate clustering results. Level set trees can also be used to construct hierarchical groupings of lattice points based on an arbitrary function evaluated on the lattice. Finally, we explore the use of level set trees for statistical inference, proposing simple methods based on repeated subsampling to describe the stochastic fluctuations in tree features. We use the new python toolbox DEnsity-BAsed CLustering (DeBaCl) to illustrate our methods on several simulated and neuroimaging data sets.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2013 program
|
2013 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Continuing Education program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.