Neural network and principal component regression in non-destructive soluble solids content assessment:a comparison

Visible and near infrared spectroscopy is a non-destructive,green,and rapid technology that can be utilized to estimate the components of interest without conditioning it,as compared with classical analytical methods.The objective of this paper is to compare the performance of artificial neural network(ANN)(a nonlinear model)and principal component regression(PCR)(a linear model)based on visible and shortwave near infrared(VIS-SWNIR)(400-1000 nm)spectra in the non-destructive soluble solids content measurement of an apple.First,we used multiplicative scattering correction to pre-process the spectral data.Second,PCR was applied to estimate the optimal number of input variables.Third,the input variables with an optimal amount were used as the inputs of both multiple linear regression and ANN models.The initial weights and the number of hidden neurons were adjusted to optimize the performance of ANN.Findings suggest that the predictive performance of ANN with two hidden neurons outperforms that of PCR.

Determination of soluble solid content and acidity of loquats based on FT-NIR spectroscopy

The near infrared （NIR） spectroscopy technique has been applied in many fields because of its advantages of simple preparation, fast response, and non-destructiveness. We investigated the potential of NIR spectroscopy in diffuse reflectance mode for determining the soluble solid content （SSC） and acidity （pH） of intact loquats. Two cultivars of loquats （Dahongpao and Jiajiaozhong） harvested from two orchards （Tangxi and Chun＇an, Zhejiang, China） were used for the measurement of NIR spectra between 800 and 2500 nm. A total of 400 loquats （100 samples of each cultivar from each orchard） were used in this study. Relationships between NIR spectra and SSC and acidity of loquats were evaluated using partial least square （PLS） method. Spectra preprocessing options included the first and second derivatives, multiple scatter correction （MSC）, and the standard normal variate （SNV）. Three separate spectral windows identified as full NIR （800-2500 nm）, short NIR （800-1100 rim）, and long NIR （1100-2500 nm） were studied in factorial combination with the preprocessing options. The models gave relatively good predictions of the SSC of loquats, with root mean square error of prediction （RMSEP） values of 1.21, 1.00, 0.965, and 1.16 °Brix for Tangxi-Dahongpao, Tangxi-Jiajiaozhong, Chun＇an-Dahongpao, and Chun＇an-Jiajiaozhong, respectively. The acidity prediction was not satisfactory, with the RMSEP of 0.382, 0.194, 0.388, and 0.361 for the above four loquats, respectively. The results indicate that NIR diffuse reflectance spectroscopy can be used to predict the SSC and acidity of loquat fruit.