Abstract
Two different electronic nose systems (metal oxide and conducting polymer based) were used to identify Salmonella typhimurium contaminated beef strip loin samples (stored at two temperatures). The sensors present in the two systems were ranked based on their Fisher criteria of ranking to evaluate their importance in discriminant analysis. The most informative sensors were then used to develop linear discriminant analysis and quadratic discriminant analysis-based classification models. Further, sensor signals collected from both the sensor systems were combined to improve the classification accuracies. The developed models classified meat samples based on the Salmonella population into "No Salmonella" (microbial counts < 0. 7 log 10 cfu/g) and "Salmonella inoculated" (microbial counts ≥ 0. 7 log 10 cfu/g). The performances of the developed models were validated using leave-1-out cross-validation. Classification accuracies of 80% and above were observed for the samples stored at 10 °C using the sensor fusion approach. However, the classification accuracies were relatively low for the meat samples stored at 4 °C when compared to the samples stored at 10 °C. The results indicate that the electronic nose systems could be effectively used as a first stage screening device to identify the meat samples contaminated with S. typhimurium.
Original language | English (US) |
---|---|
Pages (from-to) | 1206-1219 |
Number of pages | 14 |
Journal | Food and Bioprocess Technology |
Volume | 5 |
Issue number | 4 |
DOIs | |
State | Published - May 2012 |
Externally published | Yes |
Bibliographical note
Funding Information:Acknowledgements The authors would like to express their profound gratitude to the United States Department of Agriculture-Cooperative State Research, Education and Extension Service (USDA-CSREES) for their financial support for this research.
Keywords
- Discriminant analysis
- Electronic nose
- Food safety
- Intelligent sensors
- Meat contamination
- Salmonella typhimurium
- Sensor fusion