Statistical Analysis on Tree-structured objects:

1. Object Oriented Data Analysis: Sets of Trees. (with J. S. Marron)

Here are some movies for this paper: (It may take around 30 seconds to load one movie)

·         3D view of the individual simplified blood vessel trees:

first person second person and third person

·         3D view of the sample of simplified blood vessel trees: sample

·         The attribute treeline passing through the median-mean tree

·         The attribute treeline passing through the average support tree

The complete draft of the paper with detail proofs:

2. A Principal Component Analysis for Trees. (with B. Aydin, G. Pataki, E. Bullitt and J. S. Marron )

A draft of the paper: arXiv

Here are some supplemental figures for this paper:

·         The support trees of 3 sub-populations for two correspondences: ST

·         For descendant correspondence, the first three principal components for three sub-populations: Lines

·         For left sub-population, comparison of # of nodes explained with descendant and thickness correspondences: Left Comparison

The paper includes a comparison of thickness and descendant correspondence for back, left and right sub-populations:

Thickness Correspondence:

o    Scatterplots of scores belonging to principal components 1, 2, 3 and 4:

back, left and right

o    Scatterplots of *cumulative* scores belonging to principal components 1, 2, 3 and 4:

back, left and right

Descendant Correspondence:

o    Scatterplots of scores belonging to principal components 1,2,3 and 4:

back, left and right

o    Scatterplots of *cumulative* scores belonging to principal components 1, 2, 3 and 4:

back, left and right

o    Scatterplots of age parameter with principal components 1, 2 and 3:

back, left and right