Evaluating a new classification method using PCA to human activity recognition

Author

Abidine, M´hamed Bilal ; Fergani, B.

Author_Institution

Fac. of Electron. & Comput. Sci., USTHB Algiers, Algiers, Algeria

fYear

2013

fDate

20-22 Jan. 2013

Firstpage

1

Lastpage

4

Abstract

The ability to recognize human activities from sensed information is very important for ubiquitous computing applications using smart identification technologies. This paper addresses a new discriminative supervised method combining the classical Principal Components Analysis with the correlation criterion for performing activity recognition in a smart home. We conduct several experiments, demonstrate the achievement of our method and show its promising results using real world dataset.

Keywords

home computing; learning (artificial intelligence); pattern classification; principal component analysis; ubiquitous computing; PCA; classification method; discriminative supervised method; human activity recognition; principal component analysis; smart home; smart identification technology; ubiquitous computing application; Correlation; Hidden Markov models; Intelligent sensors; Principal component analysis; Smart homes; Ubiquitous computing; activity recognition; machine learning; sensors network; smart home; ubiquitous computing;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Medical Applications (ICCMA), 2013 International Conference on

Conference_Location

Sousse

Print_ISBN

978-1-4673-5213-0

Type

conf

DOI

10.1109/ICCMA.2013.6506158

Filename

6506158