170 – Issues in Building Imputation Models for Missing Data Techniques
A Review on Combination-based Tests for Shape Analysis
Chiara Brombin
University of Milan - San Raffaele
Luigi Salmaso
University of Padova
This paper reviews main features of the combination-based approach for shape analysis.
It is worth noting that inference in shape analysis is a crucial point and robust testing is much needed. Our nonparametric permutation-based approach provides a suitable and
powerful testing also in presence of many correlated landmarks and a limited sample size. An important feature of combination-based tests is the finite sample consistency property (Pesarin and Salmaso, 2010 and Brombin and Salmaso, 2013) which allows to gain power in the testing procedure by increasing the number of variables rather than the sample size, provided that added variables yield additional information.