Winter jujube (Zizyphus jujuba Mill.) quality forecasting method based on electronic nose. (1st March 2015)

• Record Type: Journal Article
• Authors: Hui, Guohua; Jin, Jiaojiao; Deng, Shanggui; Ye, Xiao; Zhao, Mengtian; Wang, Minmin; Ye, Dandan
• Is Part Of: Food chemistry. Volume 170(2015)
• Page Start: 484
• Page End: 491
• Publication Date: 2015-03-01
• Subjects: Winter jujube quality -- Electronic nose -- Double-layered cascaded series stochastic resonance -- Signal-to-noise ratio spectrum; Food -- Analysis -- Periodicals; Food -- Composition -- Periodicals; 664
• DOI: 10.1016/j.foodchem.2014.08.009
• Languages: English
• ISSNs: 0308-8146

Abstract

Highlights: Winter jujube ( Zizyphus jujuba Mill.) quality forecasting method based on electronic nose has been investigated. Physical/chemical indexes were examined to validate jujube quality changes during experiments. MVA is used to investigate the linearity relationship between SR/DCSSR eigen values and jujube physical/chemical indexes. Jujube quality forecasting model was developed based on DCSSR SNR-MAX values. The developed model presented a forecasting accuracy of 97.35%. Abstract: Winter jujube ( Zizyphus jujuba Mill.) quality forecasting method utilising electronic nose ( EN ) and double-layered cascaded series stochastic resonance (DCSSR) was investigated. EN responses to jujubes stored at room temperature were continuously measured for 8 days. Jujubes' physical/chemical indexes, such as firmness, colour, total soluble solids (TSS), and ascorbic acid (AA), were synchronously examined. Examination results indicated that jujubes were getting ripe during storage. EN measurement data was processed by stochastic resonance (SR) and DCSSR. SR and DCSSR output signal-to-noise ratio (SNR) maximums (SNR-MAX) discriminated jujubes under different storage time successfully. Multiple variable regression (MVR) results between physical/chemical indexes and SR/DCSSR eigen values demonstrated that DCSSR eigen values were more suitable for jujube quality determination. Quality forecasting model was developed using non-linear fitting regression of DCSSR eigen values. Validating experiments demonstrated that forecasting accuracy of this model is 97.35%. This method also presented other advantages including fast response, non-destructive, etc.