A novel method using MOS electronic nose and ELM for predicting postharvest quality of cherry tomato fruit treated with high pressure argon

Abstract

Flavor characteristic, caused by a series of physicochemical changes, is an important part of fruit quality and freshness. The purpose of this study was to investigate the reliability and validity of using electronic nose (e-nose) technology to evaluate the quality and freshness of cherry tomatoes after high pressure argon treatments. Extreme learning machine (ELM) and Partial least squares (PLS) were applied for quality parameters prediction. The results showed that freshness of cherry tomatoes based on flavor characteristic over the entire storage time displayed four groups (fresh, acceptable, stale and very stale, respectively). High pressure argon at 0.8 MPa significantly maintained the freshness and quality of cherry tomatoes during cold storage. ELM models showed better performance than PLS models for evaluating the firmness, SSC and pH values according to their higher fitting correlation coefficients (R2 > 0.9500). The results demonstrated e-nose technology combined with ELM provided a reliable and valid method for evaluating quality and freshness of cherry tomatoes during cold storage.