|
Activity Number:
|
420
- Contributed Poster Presentations: Social Statistics Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Social Statistics Section
|
|
Abstract #322615
|
|
|
Title:
|
Functional Principal Component Logistic Regression Model for Online Arabic Handwriting Recognition
|
|
Author(s):
|
Amani Albaqshi* and Khalil Shafie
|
|
Companies:
|
University of Northern Colorado and University of Northern Colorado
|
|
Keywords:
|
Functional data analysis ;
Principal component ;
Logistic model ;
Arabic handwriting recognition ;
Gender recognition ;
Cubic B-spline smoothing
|
|
Abstract:
|
Principal Components Analysis (PCA) is a well-known procedure that reduces dimensionality and solves multicollinearity problems in multivariate statistical methods. Functional data analysis (FDA) methodologies have developed as a natural generalization of multivariate data analysis techniques to the case where data are curves measured as functions of a continuous parameter such as time. The functional principal component regression (FPCR) method received the most attention both in aspects of the application and methodological development to improve the estimation of the functional parameter in FDA. Handwriting recognition is an active research area in pattern recognition and shapes analysis, and it provides as an infinite dimensional data object. An FPCR technique is used to classify Arabic handwriting samples taken from differential individuals. In this study, the Functional Principal Component Logistic Regression (FPC-LR) model is applied to online Arabic handwriting data with respect to gender, analyzing the word "Shm" the Arabic term for the arrow. The FPC-LR model captures a substantial amount of variance in the scripts across replication, dividing gender into writer-specific
|
Authors who are presenting talks have a * after their name.
