Abstract: This study aimed to establish different models to predict some quality parameters (soluble solids, titratable acid and moisture contents) of ‘Lingwuchangzao’ jujube based on dielectric loss factor ?’’ and dielectric constant ?’ spectra and to select the optimal one. The effective information from dielectric spectra was extracted by genetic algorithm (GA) and the correlation coefficient method. The prediction models for quality parameters were established using partial least squares (PLS), principal components regression (PCR) and support vector machine (SVM). The optimal prediction model was determined by the coefficient of determination (R2), root mean square error of calibration set (RMESC) and root mean square error of predication set (RMSEP). The results showed that the optimal prediction models for soluble solids, titratable acid and moisture contents were GA-PCR, GA-PLS and GA-PLS based on dielectric loss factor ?’’ spectra with a R2 value of greater than 0.9, respectively. The best prediction models based on dielectric constant ?’ spectra were CC-PLS, GA-SVM and GA-PLS with a R2 value of greater than 0.8, respectively. These models were validated with satisfactory results. In conclusion, this study can provide a reliable method for rapid prediction of jujube quality using dielectric spectra.

Key words: ‘Lingwuchangzao’ jujube, nondestructive detection, dielectric properties, internal quality, models