Application of MOS based electronic nose for the prediction of banana quality properties

This study deals with the examination of the ability of a low-cost electronic nose (e‐nose) for prediction of banana quality indices such as total soluble solids (TSS), titratable acidity (TA), pH and firmness at different shelf-life stages. The relationships between sensor array responses of e-nose and quality indices of banana were established by means of partial least squares (PLS), multiple linear regression (MLR) and support vector regression (SVR) techniques. All models for firmness and TSS showed a good prediction performance. However for the TA and pH, there were a poor correlation with the signal of the e-nose in MLR and PLS models. The results proved that performance of SVR models for prediction of the quality indices of banana were better than others, with high correlation coefficients of the cross validation (R² = 0.8852 for firmness, 0.9608 for TSS, 0.7607 for pH and 0.7033 for TA) and relatively low RMSE values of 1.1716 for firmness, 0.9308 for TSS, 0.1523 for pH and 0.0267 for TA. Finally, these results demonstrated that e-nose has the potential of becoming a reliable instrument to estimate chemical and physical properties of banana from signals of an e-nose system.