|
Activity Number:
|
361
- Contributed Poster Presentations: Section on Nonparametric Statistics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #312328
|
|
|
Title:
|
Discrimination of Functional Data by Detecting the Second Moment Structure of Rescaled Functions
|
|
Author(s):
|
Shuhao Jiao* and Hernando Ombao
|
|
Companies:
|
KAUST and King Abdullah Univ. of Science and Technology (KAUST)
|
|
Keywords:
|
Classification;
Dimension reduction;
Functional data analysis;
Second moment structure
|
|
Abstract:
|
This article presents a new classification method for functional data. We consider the case where two groups of functions have similar means where it would be challenging to classify based solely on the mean function. Our proposed method considers both the first and second moments. Here, we demonstrate that the new method is sensitive to divergence in the second moment structure and thus produces lower rate of misclassification compared to other competitor methods. Our method uses the Hilbert Schmidt norm to measure the divergence of second moment structure. Good performance can be achieved by dimension reduction to select the most pronounced features of discrepancy. A series of orthonormal basis is proposed and we show these basis functions account for most of the discrepancy of second moment structure. Consistency properties for the proposed estimators are established. Simulation study and real data analysis on phoneme and brain activity empirically validate the superiority of the proposed method.
|
Authors who are presenting talks have a * after their name.
