Averaging neural network ensembles model for quantification of volatile organic compound

Atiq Ur Rehman, Amine Bermak

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

After a certain period of time, there is a change in response of the gas sensors, which is caused by drift. This change in response of the gas sensors causes deterioration which makes the artificial intelligence algorithms worthless for prediction. As the algorithms are trained on data without drift and once the effect of drift starts causing an error in prediction, the system needs re-calibration, which is a cumbersome process. Neural Networks (NN) are proved to have the capability of solving many complex problems in different fields. In this paper, an averaging Neural Network ensemble model is proposed to compensate the effect of drift in gas sensors and is tested for quantification of Industrial gases. The dataset used for validating the proposed model is a large scale experimental data, available online.

Original languageEnglish
Title of host publication2019 15th International Wireless Communications and Mobile Computing Conference, IWCMC 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages848-852
Number of pages5
ISBN (Electronic)9781538677476
DOIs
Publication statusPublished - Jun 2019
Event15th IEEE International Wireless Communications and Mobile Computing Conference, IWCMC 2019 - Tangier, Morocco
Duration: 24 Jun 201928 Jun 2019

Publication series

Name2019 15th International Wireless Communications and Mobile Computing Conference, IWCMC 2019

Conference

Conference15th IEEE International Wireless Communications and Mobile Computing Conference, IWCMC 2019
Country/TerritoryMorocco
CityTangier
Period24/06/1928/06/19

Keywords

  • Concentration estimation
  • Electronic nose
  • Ensemble learning
  • Neural Networks
  • Sensors drift

Fingerprint

Dive into the research topics of 'Averaging neural network ensembles model for quantification of volatile organic compound'. Together they form a unique fingerprint.

Cite this