Optimizing genetic algorithm–partial least squares model of soluble solids content in Fukumoto navel orange based on visible–near‐infrared transmittance spectroscopy using discrete wavelet transform

Abstract

BACKGROUND The thick rind of Fukumoto navel orange is a great barrier to light penetration, which makes it difficult to evaluate the internal quality of Fukumoto navel orange accurately by visible–near‐infrared (Vis‐NIR) transmittance spectroscopy. The information carried by the transmission spectrum is limited. Thus, the application of genetic algorithm (GA) for variable selection may not reach the expected results, and selected variables may contain redundancy. In this paper, we present the use of discrete wavelet transforms for optimizing a GA–partial least squares (PLS) model based on Vis‐NIR transmission spectra of Fukumoto navel orange. Haar, Db, Sym, Coif and Bior wavelets were used to compress the spectral data selected by GA. Then a PLS model was established based on the variables compressed by each wavelet function. RESULTS The use of Db4, Sym4, Coif2 and Bior3.5 succeeded in further simplification of the GA‐PLS model by reducing the number of variables by 40–44% without decreasing the prediction accuracy. The application of Bior3.5 not only could reduce the number of variables in the GA‐PLS model by 40%, but also increase the value of correlation coefficient of prediction by 1% and decrease the value of root mean square error of prediction by 3%. CONCLUSIONS The results indicated that the combination of GA and discrete wavelet transforms for variable selection in the internal quality assessment of Fukumoto navel orange by Vis‐NIR transmittance spectroscopy was feasible. © 2019 Society of Chemical Industry

An Effective Prediction Approach for Moisture Content of Tea Leaves Based on Discrete Wavelet Transforms and Bootstrap Soft Shrinkage Algorithm

“…BOSS is a method using collinearity to select effective features and using the information of the regression coefficient to flexibly shrink the information of interest. The BOSS algorithm is constructed using bootstrap sampling (BBS) and weighted bootstrap sampling (WBS) to generate random combination of variables and sub-model, and by combining model population analysis (MPA) and the PLS algorithm to extract effective information from the sub-model [12].…”

“…The moisture content (MC), soluble solids content (SSC), pH, and hardness of Gala apple samples were tested non-destructively within 350-2500 nm using the wavelet transform pretreatment of raw spectral data [11]. The use of DWT successfully further simplified the genetic algorithm-the partial least squares (GA-PLS) model by reducing variables by 40-44% without reducing the prediction accuracy [12]. Other experimental results [13] showed that the DWT-support vector regression (DWT-SVR) multivariate regression model, having good robustness, can measure protein, starch, and fat contents in corn simultaneously, demonstrating that DWT can effectively remove noise from corn NIRS spectral data.…”

Effect of spectrum measurement position on detection of Klason lignin content of snow pears by a portable NIR spectrometer

“…In the CARS algorithm, variables with smaller absolute values of the regression coefficient were forced to be removed, but the elimination of variables may have contained useful information because of the absolute value changed with the variations of the sample space. A genetic algorithm (GA) was used to select effective variables for determination of the soluble solids concentration (SSC) in apples, pears, oranges, apricots, and garlic by Li (Li, Huang, et al, 2018), Puneet Mishra (Puneet Mishra, Woltering, Brouwer, & Hogeveen‐van Echtelt, 2021b), Song (Song et al, 2019), Ozdemir (Ozdemir et al, 2019), and Rahman (Rahman et al, 2018), respectively. Camps (Camps & Camps, 2019) used FT‐NIR spectroscopy on peeled tubers combined with a GA method to predict the reducing sugars and dry matter in potatoes.…”

Multi-product calibration model for soluble solids and water content quantification in Cucurbitaceae family, using visible/near-infrared spectroscopy

“…The Vis/NIR region (380–1050 nm) is promising as this is usually attributed to the 3rd and 4th overtones of O–H and C–H bands in sugar molecules ( Cen and He, 2007 ). Numerous studies on Vis/NIR application in SSC quantification have also been conducted on citrus fruits (P. Li et al., 2020 ; Song et al., 2019 ), apple ( Fan et al., 2020 ; Lan et al., 2021 ; Xia et al., 2020 ), pear ( Mishra et al., 2021 ), melon (M. Li et al., 2019 ), grape ( Fernández-Novales et al., 2019 ) and tomato ( Huang et al., 2018 ). This technique was also able to accurately predict water content in pomelo ( Xu et al., 2020 ), dates ( Alhamdan and Atia, 2017 ), plum ( Mulisa Bobasa et al., 2020 ; Posom et al., 2020 ), maize seed ( Zhang and Guo, 2020 ) and olives ( Lee et al., 2018 ).…”