Abstract

Abstract. In order to explore the potential of simultaneously detecting the soluble solid content (SSC) and pH of a variety of pears using visible-near infrared spectra technology, linear and nonlinear predictive models for SSC and pH of three kinds of pears were built using the partial least squares (PLS) and least squares-support vector machine (LS-SVM) methods in this study. One-hundred eighty samples (60 for each variety) were selected as sample set. 150 pear samples (50 for each variety) were selected randomly for the calibration set, and the remaining 30 samples (10 for each variety) for the prediction set. Different preprocessing methods including average smoothing, first and second derivatives were applied to improve the predictive ability of the models. For estimation of SSC and pH values, PLS models with original spectra and LS-SVM models with the first derivative obtained the best performance. Moreover, with a comparation between PLS and LS-SVM models, it was found that PLS model based on original spectra was better than LS-SVM model. The correlation coefficients (r) of SSC and pH were 0.9205 and 0.9318, respectively. The root mean square error of prediction (RMSEP) of SSC and pH were 0.4091 and 0.0729, respectively. Test results showed that it is feasible to detect the SSC and pH of a variety of pears using visible-near infrared spectra technology simultaneously.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call