Support vector machines for inhabitant identification in smart houses

Authors:
Rachid Kadouche;Hélène Pigot;Bessam Abdulrazak;Sylvain Giroux
Affiliations:
DOMUS Lab, Université de Sherbrooke, Sherbrooke, Québec, Canada;DOMUS Lab, Université de Sherbrooke, Sherbrooke, Québec, Canada;DOMUS Lab, Université de Sherbrooke, Sherbrooke, Québec, Canada;DOMUS Lab, Université de Sherbrooke, Sherbrooke, Québec, Canada
Venue:
UIC'10 Proceedings of the 7th international conference on Ubiquitous intelligence and computing
Year:
2010

Citing 9
Cited 1

Classification by pairwise coupling

NIPS '97 Proceedings of the 1997 conference on Advances in neural information processing systems 10
Biometric identification

Communications of the ACM
Gene Selection for Cancer Classification using Support Vector Machines

Machine Learning
The smart floor: a mechanism for natural user identification and tracking

CHI '00 Extended Abstracts on Human Factors in Computing Systems
Gait-Based Recognition of Humans Using Continuous HMMs

FGR '02 Proceedings of the Fifth IEEE International Conference on Automatic Face and Gesture Recognition
Variable selection using svm based criteria

The Journal of Machine Learning Research
Kernel Methods for Pattern Analysis

Kernel Methods for Pattern Analysis
Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)

Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)
An introduction to biometric recognition

IEEE Transactions on Circuits and Systems for Video Technology

A non-intrusive monitoring system for ambient assisted living service delivery

ICOST'12 Proceedings of the 10th international smart homes and health telematics conference on Impact Ananlysis of Solutions for Chronic Disease Prevention and Management

Quantified Score

Hi-index	0.00

Visualization

Abstract

Authentication is the process by which a user establishes his identification when accessing a service. The use of password to identify the user has been a successful technique in conventional computers. However, in pervasive computing where computing resources exist everywhere, it is necessary to perform user identification through various means. This paper addresses the inhabitant identification issue in smart houses. It studies the optimum time and sensor set required to unobtrusively detect the house occupant. We use a supervised learning approach to address this issue by learning Support Vector Machines classifier (SVM), which predict the users by their daily life habits. We have analyzed the early morning routine with six users. From the very first minute, users can be recognized with an accuracy of more than 85%. Then we have applied an SVM feature selection algorithm to remove noisy and outlier features. Thus, this increases the accuracy to 88% using less then 10 sensors.