Gas sensor drift mitigation using classifier ensembles

Authors:
Alexander Vergara;Ramón Huerta;Tuba Ayhan;Margaret Ryan;Shankar Vembu;Margie Homer
Affiliations:
University of California, San Diego, La Jolla, CA;University of California, San Diego, La Jolla, CA;Technical University of Istanbul, Maslak, Istanbul, Turkey;California Institute of Technology, Pasadena, CA;University of California, San Diego, La Jolla, CA;California Institute of Technology, Pasadena, CA
Venue:
Proceedings of the Fifth International Workshop on Knowledge Discovery from Sensor Data
Year:
2011

Citing 9
Cited 0

Support-Vector Networks

Machine Learning
Nonlinear component analysis as a kernel eigenvalue problem

Neural Computation
Dynamic Weighted Majority: A New Ensemble Method for Tracking Concept Drift

ICDM '03 Proceedings of the Third IEEE International Conference on Data Mining
Mining concept-drifting data streams using ensemble classifiers

Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining
In Defense of One-Vs-All Classification

The Journal of Machine Learning Research
Using additive expert ensembles to cope with concept drift

ICML '05 Proceedings of the 22nd international conference on Machine learning
The Forgetron: A Kernel-Based Perceptron on a Budget

SIAM Journal on Computing
Dynamic Weighted Majority: An Ensemble Method for Drifting Concepts

The Journal of Machine Learning Research
LIBSVM: A library for support vector machines

ACM Transactions on Intelligent Systems and Technology (TIST)

Quantified Score

Hi-index	0.00

Visualization

Abstract

Sensor drift remains to be one of the challenging problems in chemical sensing. To address this problem we collected an extensive data set for six different volatile organic compounds over a period of three years under tightly-controlled operating conditions using an array of 16 metal-oxide sensors. We then adopted a machine learning approach namely an ensemble of classifiers to cope with sensor drift. For this particular application we chose support vector machine as our base classifier in the ensemble but, in principle, any other classifier can be used. Experiments clearly indicate the presence of drift in the sensors during the period of three years and that it degrades the performance of classifiers. However, the ensemble method that uses a weighted combination of classifiers trained at different points of time is able to cope well with sensor drift.