Classification of Three Volatiles Using a Single-Type eNose with Detailed Class-Map Visualization
The use of electronic noses (eNoses) as analysis tools are growing in popularity; however, the lack of a comprehensive, visual representation of how the different classes are organized and distributed largely complicates the interpretation of the classification results, thus reducing their practical...
| Autores: | , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | España |
| Recursos: | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| Repositorio: | Recercat. Dipósit de la Recerca de Catalunya |
| OAI Identifier: | oai:recercat.cat:10459.1/84134 |
| Acesso em linha: | https://doi.org/10.3390/s22145262 http://hdl.handle.net/10459.1/84134 |
| Access Level: | acceso abierto |
| Palavra-chave: | Electronic nose eNose Array of gas sensors MOX gas sensors PCA and LDA analysis |
| Resumo: | The use of electronic noses (eNoses) as analysis tools are growing in popularity; however, the lack of a comprehensive, visual representation of how the different classes are organized and distributed largely complicates the interpretation of the classification results, thus reducing their practicality. The new contributions of this paper are the assessment of the multivariate classification performance of a custom, low-cost eNose composed of 16 single-type (identical) MOX gas sensors for the classification of three volatiles, along with a proposal to improve the visual interpretation of the classification results by means of generating a detailed 2D class-map representation based on the inverse of the orthogonal linear transformation obtained from a PCA and LDA analysis. The results showed that this single-type eNose implementation was able to perform multivariate classification, while the class-map visualization summarized the learned features and how these features may affect the performance of the classification, simplifying the interpretation and understanding of the eNose results. |
|---|