No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Determination of Internal Qualities of Newhall Navel Oranges Based on NIR Spectroscopy Using Machine Learning
... The MSC-PLSR model exhibited the best performance among all of the constructed models for Vc predictions. These results are consistent with those of Liu et al. [39]. The predictive capacity of the surface color indexes (L*, a*, and b*) is presented in Table 2. ...
... According to the literature [39], an RPD should be greater than 2.0 for reasonable predictions, while a value of over 3.0 indicates reliable predictions. Thus, the models proposed in this study for weight loss rate and L* can be utilized for analytical purposes, and the prediction models for all attributes can be considered to be reliable. ...
... Thus, the models proposed in this study for weight loss rate and L* can be utilized for analytical purposes, and the prediction models for all attributes can be considered to be reliable. Moreover, SVM performed consistently better than PLS in predicting total soluble solids, Vc content, and titratable acid [39], which indicates that SVM is more capable of handling noisy and high-dimensional data. However, in this study, there was inefficient evidence to determine whether PLSR performed better or worse than SVR for quality attributes in mini-Chinese cabbage with nanopacking. ...
A nondestructive optical method is described for the quality assessment of mini-Chinese cabbage with nanopackaging during its storage, using Fourier transform-near infrared (FT-NIR) spectroscopy. The sample quality attributes measured included weight loss rate, surface color index, vitamin C content, and firmness. The level of freshness of the mini-Chinese cabbage during storage was divided into three categories. Partial least squares regression (PLSR) and the least squares support vector machine were applied to spectral datasets in order to develop prediction models for each quality attribute. For a comparative analysis of performance, the five preprocessing methods applied were standard normal variable (SNV), first derivative (lst), second derivative (2nd), multiplicative scattering correction (MSC), and auto scale. The SNV-PLSR model exhibited the best prediction performance for weight loss rate (Rp2 = 0.96, RMSEP = 1.432%). The 1st-PLSR model showed the best prediction performance for L* value (Rp2 = 0.89, RMSEP = 3.25 mg/100 g), but also the lowest accuracy for firmness (Rp2 = 0.60, RMSEP = 2.453). The best classification model was able to predict freshness levels with 88.8% accuracy in mini-Chinese cabbage by supported vector classification (SVC). This study illustrates that the spectral profile obtained by FT-NIR spectroscopy could potentially be implemented for integral assessments of the internal and external quality attributes of mini-Chinese cabbage with nanopacking during storage.
... Fu et al. (2008) found that the spectra acquired from different latitudes had greater difference than the spectra from different longitudes in pear [16]. Liu et al. (2015) found the model established with the spectra of equator region was slight better than calyx region in Newhall navel orange [17]. Han et al. (2015) also found that the vertical upward orientation of stem-calyx axis was more suitable for SSC evaluation than other orientations in apple [18]. ...
... Fu et al. (2008) found that the spectra acquired from different latitudes had greater difference than the spectra from different longitudes in pear [16]. Liu et al. (2015) found the model established with the spectra of equator region was slight better than calyx region in Newhall navel orange [17]. Han et al. (2015) also found that the vertical upward orientation of stem-calyx axis was more suitable for SSC evaluation than other orientations in apple [18]. ...
... Difference of texture and character among citrus varieties were extremely remarkable, which may influence prediction accuracy of Vis/NIR on fruit quality parameters [29,30,31]. At present, more researches in the field of nondestructive testing focus on mandarin [10,[31][32][33][34][35][36] and orange [17,23,29,[37][38][39][40], but few focus on the endemic species. The objective of this study was to develop a accurate estimation model for SSC based on the spectra collected from multiple regions of intact citrus. ...
Fruit is complicated natural production and the chemical components are heterogeneous at different position, the comprehensive quality evaluation for intact fruit is the fairest way for avoiding the nonuniform distribution. In order to build a higher accuracy model for soluble solids content (SSC) evaluation, the visible/near infrared spectra of citrus were collected from stem, equator and navel positions to analyze the influence of spectrum measurement position on the prediction accuracy of SSC. The SSC value gradually reduced from navel to equator and stem positions by sequence, but the trend of peel thickness of these three regions were opposite of SSC. The coefficient correlation of SSC between intact fruit and three local regions reached remarkable level, which proved it was feasible to build an accurate model for SSC prediction of intact fruit using the spectral information of local positions. Then separate local models based on specific position spectrum (stem, equator and navel) were built, the result showed the equator position was more suitable to evaluate SSC of intact fruit than navel and stem positions due to the better prediction accuracy, however the unpredictability and variation of the spectral collection position is a challenge to the prediction ability of intact fruit quality. Next multi-region combination models that fusing spectral information of multiple positions were developed, the combination models of ‘Equator + Navel’ and ‘Stem + Equator + Navel’ achieved optimal performance than other combination models and all separate local models, with the correlation coefficient of prediction set (Rpre) and root mean square error of prediction (RMSEP) of 0.8507, 0.8424 and 0.6015°Brix, 0.5901°Brix, respectively. It indicated that the peel thickness interfered the acquisition of spectral information of flesh layer, but the accuracy and robust of the prediction model could be improved through fusing the spectral information of multiple regions. Therefore, the fusion of multi-information sets should be deserved more attention to build a practicable model that is not sensitive to the variation of spectrum measurement position for SSC evaluation of intact citrus fruit.
... Vis/NIRS, together with chemometric softwares, has been successfully applied for prediction of different parameters on various citrus fruit in previous studies (Peiris et al., 1998;Tsuchikawa et al., 2003;Lee et al., 2004;Kim et al., 2004;Gomez et al., 2006;Zheng et al., 2010;Liao et al., 2013;Sànchez et al., 2013;Magwaza et al., 2014;Liu et al., 2015). Physio-chemical properties such as rind dry matter (DM), non-structural carbohydrates and antioxidants content, that might be linked with susceptibility of 'Nules Clementine' mandarin fruit to rind breakdown disorder, can be predicted using Vis/NIRS Magwaza et al., 2014a, b). ...
... Citrus biochemical compounds are characterized by bonds such as O-H, N-H, C-O and C-H that can be studied by illuminating the fruit with radiation and investigating light transmission, reflection and refraction pattern (Liu et al., 2015). The introduction of near infrared radiation application to study intact biological samples has necessitated the investigation of internal and external quality of fruit without causing any damage to the fruit . ...
... Near infrared spectroscopy (NIRS) is arguably the latest and most advanced technology used for such an objective. Previous studies of using NIRS on citrus have shown the technique to be appropriate and successful with relevant chemometrics to analyze spectra (Liao et al., 2013;Sànchez et al., 2013;2014b, c;Liu et al., 2015). This paper aims to discuss current knowledge about CI, RP, and application of NIRS to predict quality parameters and physiological disorders of citrus fruit. ...
In this study, visible to near infrared spectroscopy (Vis/NIRS) was investigated as a tool for non-destructive prediction of ‘Marsh’ grapefruit susceptibility to postharvest rind physiological disorders, specifically chilling injury and rind pitting. However, before such objective techniques were developed, specific pre-symptomatic biochemical markers related to chilling injury and rind pitting disorders were identified. After which, the potential to predict these biomarkers and susceptibility of fruit to these disorders was investigated. The first chapter is a general introduction outlining aims and objectives of this research. The second chapter reviews previous literature in an attempt to gain an understanding of the biological mechanisms for citrus rind physiological disorders. This chapter also reviews the applications of Vis/NIRS for non-destructive prediction of citrus fruit quality. The third chapter is evaluating the conditions and potential of using Vis/NIRS to estimate internal citrus fruit parameters. The study in chapter 3 was out of ‘Marsh’ grapefruit harvest season and therefore, ‘Star Ruby’ grapefruit and ‘Valencia’ orange were used instead of ‘Marsh’ because they possess similar characteristics. The fourth chapter is identifying physiological attributes that can be used as pre-symptomatic markers during prediction of ‘Marsh’ grapefruit postharvest rind disorders. The fifth chapter is the prediction of ‘Marsh’ grapefruit physiological rind disorders and their pre-symptomatic biochemical markers. Overall discussions and conclusions are made in chapter six, where future research prospects are also recommended.
... During the entire ripening process, grapes undergo biochemical and physiological changes, i.e., softening, pigment accumulation, and flavor and aroma formation [5]. For these changes, ripeness is often evaluated indirectly based on parameters such as soluble solids content (SSC) and total acid (TA) levels, as they reflect grape quality and ripeness, with low SSC and high TA values indicating unripe and less flavorful grapes [6][7][8]. However, the traditional method of testing these parameters is predicated on destroying the integrity of the whole grape bunch and requires a large number of samples, which is time consuming and uneconomical, as well as requires professional operators with expertise [9]. ...
... In spectral analysis, it is an important step to preprocess the obtained spectral data with appropriate preprocessing methods to enhance the accuracy of prediction models [6]. There are six common preprocessing methods, including standard normal variate (SNV), multiple scattering correction (MSC), 1st derivative, 2nd derivative, S-G smoothing, and S-G smoothing + 1st derivative, which were applied to deal with the spectral data and helped us choose the best one. ...
Ripeness significantly affects the commercial values and sales of fruits. In order to monitor the change in grapes’ quality parameters during ripening, a rapid and nondestructive method of visible–near-infrared spectral (Vis-NIR) technology was utilized in this study. Firstly, the physicochemical properties of grapes in four different ripening stages were explored, with increasing color in redness/greenness (a*) and Chroma (C*) as well as soluble solids (SSC) content as ripening advanced, and decreasing values in color of lightness (L*), yellowness/blueness (b*), hue angle (h*), hardness, and total acid (TA) content. Based on these results, spectral prediction models for SSC and TA in grapes were established. Effective wavelengths were selected using the competitive adaptive weighting algorithm (CARS), and six common preprocessing methods were applied to pretreat the spectral data. Partial least squares regression (PLSR) was applied to establish models on the basis of effective wavelengths and full spectra. The models of SSC and TA with full spectral data using PLSR after 1st derivative preprocessing both obtained the best results. For SSC, these yielded optimum results when the coefficients of determination of PLSR models for the calibration set (RCal2) and the prediction set (RPre2) were 0.97 and 0.93, respectively; the root mean square error for the calibration set (RMSEC) and the prediction set (RMSEP) were 0.62% and 1.27%, respectively; and the RPD was 4.09. As for TA, the optimum values of RCal2, RPre2, RMSEC, RMSEP, and RPD were 0.97, 0.94, 0.88 g/L, 1.96 g/L, and 4.55, respectively. The results indicated that Vis-NIR spectroscopy is an effective tool for the rapid and non-destructive detection of SSC and TA in grapes.
... As such, infrared spectroscopy methods have been used to process simultaneous determination of several B-groups vitamins in various matrixes (Wojciechowski et al., 1998;Xiao et al., 2012). Similarly, the technique has allowed for precise quantification of ascorbic acid contents in several fruits and fruit juices (Alamar et al., 2016;Andrianjaka-Camps et al., 2015;Arendse et al., 2018;Blanco-Díaz et al., 2014;Liu et al., 2015;Yang & Irudayaraj, 2002), and therefore, appears promising for extension toward the grape must and wine matrixes. Such an analysis could either be performed in the NIR-visible region of the spectrum, at wavelengths between 400 and 2500 nm (Alamar et al., 2016;Blanco-Díaz et al., 2014;Caramês et al., 2017;Liu et al., 2015), or at MIR wavelengths, comprised between 2000 and 10,800 nm (Andrianjaka-Camps et al., 2015;Arendse et al., 2018). ...
... Similarly, the technique has allowed for precise quantification of ascorbic acid contents in several fruits and fruit juices (Alamar et al., 2016;Andrianjaka-Camps et al., 2015;Arendse et al., 2018;Blanco-Díaz et al., 2014;Liu et al., 2015;Yang & Irudayaraj, 2002), and therefore, appears promising for extension toward the grape must and wine matrixes. Such an analysis could either be performed in the NIR-visible region of the spectrum, at wavelengths between 400 and 2500 nm (Alamar et al., 2016;Blanco-Díaz et al., 2014;Caramês et al., 2017;Liu et al., 2015), or at MIR wavelengths, comprised between 2000 and 10,800 nm (Andrianjaka-Camps et al., 2015;Arendse et al., 2018). ...
Vitamins are essential compounds to yeasts, and notably in winemaking contexts. Vitamins are involved in numerous yeast metabolic pathways, including those of amino acids, fatty acids, and alcohols, which suggests their notable implication in fermentation courses, as well as in the development of aromatic compounds in wines. Although they are major components in the course of those microbial processes, their significance and impact have not been extensively studied in the context of winemaking and wine products, as most of the studies focusing on the subject in the past decades have relied on relatively insensitive and imprecise analytical methods. Therefore, this review provides an extensive overview of the current knowledge regarding the impacts of vitamins on grape must fermentations, wine-related yeast metabolisms, and requirements, as well as on the profile of wine sensory characteristics. We also highlight the methodologies and techniques developed over time to perform vitamin analysis in wines, and assess the importance of precisely defining the role played by vitamins in winemaking processes, to ensure finer control of the fermentation courses and product characteristics in a highly complex matrix.
... These techniques should be applicable in the farm for ease of assessing many representative samples in short time and in situ categorization of fruit into relevant batches. This could be achieved by application of non-destructive techniques such as radiation spectroscopy because of its ability to portray a biochemical profile of a biological sample within few seconds (Cayuela, 2008;Sánchez et al., 2013;Liu et al., 2015). Non-destructive techniques have gained interest from researchers for determining internal parameters of fruits. ...
... The reports then become unnecessary from the application point of view. As long as the ordinary PLS model or its modified forms are able to obtain 97% prediction accuracy on analysing TSS(Liu et al. 2015;Ncama et al. 2016), they are better than using the destructive reflectometer technique. As long as PLS models can obtain 90% accuracy on analysing total phenolic compounds(Mora-Ruiz et al. 2017, Genisheva et al. 2018, they are better than the use of procedures based on protocols involving use of chemicals. ...
Abstract (English)
The South African fruit industry holds a reputable record in exports of fresh fruit. However, commercial farms are currently using destructive techniques to estimate parameters such as mesocarp moisture content of avocado and juice total soluble solutes (TSS) of orange to estimate the fruits harvest maturity. These destructive techniques are laborious and results to waste of fruit during the sampling process. Most studies that demonstrate application of visible to near infrared spectroscopy (Vis-NIRS) as an accurate non-destructive technique to assess physiological maturity of fruit have been carried out in laboratories. The efficacy of models developed in controlled environments may not be guaranteed for application in the field since the physiology of hanging fruit changes with ambient conditions. In this study, optimum calibration techniques for Vis-NIRS to assess physiological maturity of on-tree fruit were investigated. There were non-significant differences (p > 0.05) between destructive and non-destructively assessed oil and moisture content of avocado, and TSS and titratable acidity of ‘Valencia’ orange. These results demonstrated the efficacy of Vis-NIRS for indexing maturity of hanging fruits. On the other hand, natural treatments of fruit get higher attention than chemical fungicides because chemical fungicides pose residual threats to consumer’s health, are expensive, and cause environmental pollution. In this study, grapefruit seed extracts (GSE) and moringa leaf extracts (MLE) were assessed for enriching activities of organic coatings. Hydrous extracts and extracts obtained using 50%, 80% and absolute (100%) ethanol were tested for antifungal activities and maintaining antioxidant compounds of grapefruit peels. Hydrous GSE or 100% ethanol extracts completely inhibited mycelial growth of Penicillium digitatum, in an in-vitro experiment. The hydrous GSE or MLE were recommended as cheap natural agents for controlling citrus green moulds. As such, the hydrous extracts were used in formulation of organic coatings for grapefruit. The difference in physiology of grapefruit coated with organic coatings or the commercially applied Citrashine® wax was not significant. Grapefruit coated with GSE or MLE had higher rating based on external appearance compared to fruit coated with carboxymethylcellulose or Citrashine®. This study demonstrated the novelty of using GSE and MLE as ingredients of edible, economic and effective organic coatings of grapefruit.
Ngokufinqiwe (isiZulu)
Imboni yezithelo zaseNingizimu Afrika inomlando ohloniphekile ekuthengisweni kwezithelo. Kodwa-ke, amapulazi ezentengiselwano asebenzisa amasu alimaza izithelo ukukala izici ezifana nezinga lomswakama otholakala ku-mesocarp yakotapheya noma izinga likashukela otholakala ema-wolinshini ukuze kulinganiselwe isikhathi esilungele ukuvuna lezi zithelo. Lezi zindlela zithatha isikhathi eside futhi zidala ukulahlekelwa izithelo ezisetshenziswa ngesikhathi socwaningo lamasampula. Ucwaningo oluningi oluveza ukusetshenziswa kwesibonakaliso se-visible to near infrared spectroscopy (Vis-NIRS) njengendlela ecokeme yokuhlola ukuvuthwa kwezithelo ngaphandle kokuzilimaza lujwayele ukwenzelwa ema-laboratories. Ukusebenza kwamamodeli e-Vis-NIRS akhelwe ezindaweni ezilawulwayo akuqinisekiselwe ukusetshenziswa ngaphandle ngoba i-physiology yezithelo iyashintshwa ukushisa kwelanga ngenxa yokushintsha kwesikhathi sosuku kanye nesimo sezulu. Kulolu cwaningo kuphenywe izimo i-Vis-NIRS ekumele ibe kuzo ukuze ikwazi ukuhlola ukuvuthwa kwezithelo zisalenga esihlahleni. Kubekhona umehluko ongabalulekile (p> 0.05) phakathi kwamafutha kanye nomswakama kukwatapheya, kanye noshukela nobumuncu be-olintshi okalwe ngokulimaza noma ngokusebenzisa i-Vis-NIRS. Lemiphumela iveze i-Vis-NIRS njengethuluzi elikulungele ukukala izinga lokuvuthwa kwezithelo ngaphandle kokuzivuna. Ngakolunye uhlangothi, ukwelashwa kwezithelo ngokwemvelo kunesasasa eliphakeme kunokusetshenziswa kwamakhemikhali ngoba amakhemikhali anomthelela ongemuhle empilweni yomthengi, ayabiza, futhi abangela ukungcoliswa kwemvelo. Kulolu cwaningo, umhluzi wezinhlamvu zobhamubhamu (GSE) nowamaqabunga e-moringa (MLE) uhlolelwe ukulungela ukwakha umuthi ophephile wokugcoba izithelo. Imihluzi beyilungiselelwe ngamanzi, ngotshwala obujiye ngo 50% nango 80%, kanye notshwala obungaxaxiwe (100% ethanol). Lemihluzi ibese yacwaningelwa ukuvimba isikhunta kanye nokulondoloza amasosha ekhasi lobhamubhamu. I-GSE ehluzwe ngamanzi noma nge-100% ethanol igweme ukukhula kwe-mycelial ye-Penicillium digitatum ngokuphelele ngenkathi kuhlolwa ngendlela ye-in vitro. I-GSE noma i-MLE ehluzwe ngamanzi iphakanyisiwe njengezakhi zendalo ezishibhile zokulawula isikhunta esiluhlaza se-citrus. Ngenxa yalokho, kusetshenziswe umhluzi wamanzi ukwakha umuthi wokugcoba ubhamubhamu. Umehluko emzimbeni wabhamubhamu ogcotshwe ngemithi eyakhiwe nge-GSE noma ekusetshenziswe umovu we-Citrashine® ovame ukusetshenziswa ezimakethe awuzange ube mukhulu ngokubalulekile. Ubhamubhamu ogcotshwe nge-GSE noma nge-MLE ukhethwe njengomuhle ekubukekeni kwengaphandle uma uqhathaniswa nogcotshwe nge-carboxymethylcellulose noma nge-Citrashine®. Kulolucwaningo kuvezwe i-GSE ne-MLE njengezakhi zomuthi ophephele ukudliwa, oshibhile futhi osebenza ngezinga elicokeme wokugcoba ubhamubhamu.
... Machine learning aims to train a predictive model from a training set of input-out pairs and then use the model to predict an unknown output from a given test input [1,13,14,17,31,34,37]. In this paper, we focus on the machine learning problem of binary pattern classification. ...
... With the optimization in the previous section, we summarize the iterative algorithm to optimize the problem in (13). The iterative algorithm is given in Algorithm 1. ...
Traditional machine learning methods usually minimize a simple loss function
to learn a predictive model, and then use a complex performance measure to
measure the prediction performance. However, minimizing a simple loss function
cannot guarantee that an optimal performance. In this paper, we study the
problem of optimizing the complex performance measure directly to obtain a
predictive model. We proposed to construct a maximum likelihood model for this
problem, and to learn the model parameter, we minimize a com- plex loss
function corresponding to the desired complex performance measure. To optimize
the loss function, we approximate the upper bound of the complex loss. We also
propose impose the sparsity to the model parameter to obtain a sparse model. An
objective is constructed by combining the upper bound of the loss function and
the sparsity of the model parameter, and we develop an iterative algorithm to
minimize it by using the fast iterative shrinkage- thresholding algorithm
framework. The experiments on optimization on three different complex
performance measures, including F-score, receiver operating characteristic
curve, and recall precision curve break even point, over three real-world
applications, aircraft event recognition of civil aviation safety, in- trusion
detection in wireless mesh networks, and image classification, show the
advantages of the proposed method over state-of-the-art methods.
... Measuring the internal quality of citrus fruit during ripeness can be challenging due to the interference of NIR light caused by the thickness of the peel . However, studies conducted on other citrus fruit, such as oranges Tian et al., 2020;Song et al., 2019;Cavaco et al., 2018;Liu et al., 2015;Sánchez et al., 2013a;Jamshidi et al., 2012;Liu et al., 2012;Cayuela , 2008;Zude et al., 2008), mandarins (Pires et al., 2022;Li et al., 2020;Torres et al., 2019;Sánchez et al., 2013b;Antonucci et al., 2011, Xudong et al., 2009Guthrie et al., 2005) and grapefruit (Ncama et al., 2017) showed positive results in prediction of quality parameters such as TSS and TA using NIRS. ...
The lemon industry has the challenge of providing fruits with high-quality standards worldwide. Replacing the subjective fruit quality assessment methods with objective and non-destructive techniques. Total soluble solids (TSS) and titratable acidity (TA) have been revealed as important ripening markers in lemons. Therefore, this study proposes, for the first time, using near-infra-red spectroscopy (NIRS) as a rapid and non-destructive alternative to evaluate these quality traits in 'Fino' lemons (Citrus limon L. Burm) during ripeness. NIR spectra (950-1700 nm) of intact lemons collected from two different orchards at three ripening stages were acquired, while standard destructive methods were used to determine TSS and TA in the juice of each fruit. The prediction of the quality parameters was carried out using partial least squares regression (PLS-R) models. Three approaches were followed to validate the models: internal, external, and recalibrated external validation. The results following the first approach presented a good predictive performance for both quality parameters (TSS: R 2 = 0.84, RMSEP = 0.42 and RPD = 2.5; TA: R 2 = 0.72, RMSEP = 0.45 and RPD = 2.0). When the external validation was performed, the best results were obtained for the TSS prediction using recalibrated models, maintaining good predictive performance accuracy (R 2 = 0.74 and 0.67, RMSEP = 0.42 and 0.58, and RPD = 2.4 and 1.7). Regarding distinguishing different origins, models based on partial least squares discriminant analysis (PLS-DA) were externally validated, achieving 66.4% correct classification, respectively. Thus, applying NIR technology in the lemon fruit packinghouses is a promising alternative to improve fruit management and meet consumer demands.
... Thus, the light scattering effect consists of both additive and multiplicative effects [9]. The tral analysis [20]. Due to the different spectral acquisition methods and data types, there may be differences in the modeling results using linear and nonlinear modeling methods. ...
Soluble solids content (SSC) is one of the main quality indicators of apples, and it is important to improve the precision of online SSC detection of whole apple fruit. Therefore, the spectral pre-processing method of spectral-to-spectral ratio (S/S), as well as multiple characteristic wavelength member model fusion (MCMF) and characteristic wavelength and non-characteristic wavelength member model fusion (CNCMF) methods, were proposed for improving the detection performance of apple whole fruit SSC by diffuse reflection (DR), diffuse transmission (DT) and full transmission (FT) spectra. The modeling analysis showed that the S/S- partial least squares regression models for all three mode spectra had high prediction performance. After competitive adaptive reweighted sampling characteristic wavelength screening, the prediction performance of all three model spectra was improved. The particle swarm optimization–extreme learning machine models of MCMF and CNCMF had the most significant enhancement effect and could make all three mode spectra have high prediction performance. DR, DT, and FT spectra all had some prediction ability for apple whole fruit SSC, with FT spectra having the strongest prediction ability, followed by DT spectra. This study is of great significance and value for improving the accuracy of the online detection model of apple whole fruit SSC.
... Machine learning thus has a close relation with the fields of statistics and data mining. NIR spectral data is very much noisy and the useful information gets suppressed in it, which needs an application of proper pre-processing techniques to improve the predictive ability of the machine learning techniques (Liu et al. 2015). NIR spectroscopy along with machine learning techniques has been extensively used in adulteration detection for a huge range of food products, like, milk powders (Mauer et al. 2009) and cocoa powders (Vasconezet al. 2018). ...
Machine learning techniques were applied systematically to the spectral data of near-infrared (NIR) spec-troscopy to find out the sudan dye I adulterants in turmeric powders. Turmeric powder is one of the most commonly used spice and a simple target for adulteration. Pure tur-meric powder was prepared at the laboratory and spiked with sudan dye I adulterants. The spectral data of these adulterated mixtures were obtained by NIR spectrometer and investigated accordingly. The concentrations of the adulterants were 1%, 5%, 10%, 15%, 20%, 25%, 30% (w/w) respectively. Exploratory data analysis was done for the visualization of the adulterant classes by principal component analysis (PCA). Optimization of the pre-processing and wavelength selection was done by cross-validation techniques using a partial least squares regression (PLSR) model. For quantitative analysis four different regression techniques were applied namely ensemble tree regression (ENTR), support vector regression (SVR), principal component regression (PCR), and PLSR, and a comparative analysis was done. The best method was found to be PLSR. The accuracy of the PLSR analysis was determined with the coefficients of determination (R 2) of greater than 0.97 and with root mean square error (RMSE) of less than 0.93 respectively. For the verification of the robustness of the model, the Figure of merit (FOM) of the model was derived with the help of the Net analyte signal (NAS) theory. The current study established that the NIR spectroscopy can be applied to detect and quantify the amount of sudan dye I adulterants added to the turmeric powders with satisfactory accuracy.
... Machine learning thus has a close relation with the fields of statistics and data mining. NIR spectral data is very much noisy and the useful information gets suppressed in it, which needs an application of proper pre-processing techniques to improve the predictive ability of the machine learning techniques (Liu et al. 2015). NIR spectroscopy along with machine learning techniques has been extensively used in adulteration detection for a huge range of food products, like, milk powders (Mauer et al. 2009) and cocoa powders (Vasconezet al. 2018). ...
Machine learning techniques were applied systematically to the spectral data of near-infrared (NIR) spectroscopy to fnd out the sudan dye I adulterants in turmeric
powders. Turmeric powder is one of the most commonly
used spice and a simple target for adulteration. Pure turmeric powder was prepared at the laboratory and spiked with
sudan dye I adulterants. The spectral data of these adulterated mixtures were obtained by NIR spectrometer and investigated accordingly. The concentrations of the adulterants
were 1%, 5%, 10%, 15%, 20%, 25%, 30% (w/w) respectively.
Exploratory data analysis was done for the visualization
of the adulterant classes by principal component analysis
(PCA). Optimization of the pre-processing and wavelength
selection was done by cross-validation techniques using a
partial least squares regression (PLSR) model. For quantitative analysis four diferent regression techniques were
applied namely ensemble tree regression (ENTR), support
vector regression (SVR), principal component regression
(PCR), and PLSR, and a comparative analysis was done.
The best method was found to be PLSR. The accuracy of
the PLSR analysis was determined with the coefcients of
determination (R2
) of greater than 0.97 and with root mean
square error (RMSE) of less than 0.93 respectively. For the
verifcation of the robustness of the model, the Figure of
merit (FOM) of the model was derived with the help of the
Net analyte signal (NAS) theory. The current study established that the NIR spectroscopy can be applied to detect and
quantify the amount of sudan dye I adulterants added to the
turmeric powders with satisfactory accuracy
... Dirt was removed from the surface of the apples and the test was performed after standing at room temperature for 24 h to eliminate the effect of temperature on spectra acquisition. All samples were divided into four positions along the equator for labeling (Liu et al., 2015). ...
In this paper, 3 years (2020, 2021, and 2022) of Bingtangxin apples were used to study the upgrade and maintenance method of apple soluble solids content (SSC) model using near infrared (NIR) spectroscopy. Modeling by partial least squares (PLS) and one‐dimensional convolutional neural network (1D‐CNN) algorithm. The “base model” is built with apples in 2020 and then upgrade and maintain it. The upgraded model is used to predict the sample in 2021 and 2022. Three methods were used to upgrade and maintenance the model, which are based on updated data, direct correction‐based method, and similar band‐based method. Among the three upgrade methods, the model maintenance method using updated data had the greatest improvement in prediction correlation coefficients (Rp) in both years, and both obtained good predictions. The use of direct correction resulted in the greatest reduction in the root mean square error of prediction (RMSEP) in both years. However, the Rp of the model is not improved much by using the direct correction method. The model maintenance method using similar bands combined with different modeling methods also gives good prediction results. Finally, after comparison, the model maintenance method of selecting similar bands is preferred. In the case of limited improvement in the effect of using the similar band method, the model is maintained using the method of adding updated data. The investigation of model upgrade and maintenance methods can reduce human and material consumption and make apple SSC models more versatile.
Practical applications
NIR technology is the first choice for online fruit quality inspection, but its inspection process requires constant upgrading and maintenance with models. The quality of apples varies from year to year, and the model built with samples from a single year does not work well in predicting samples from other years, resulting in the need to upgrade and maintain the fruit quality detection model. However, sometimes, although a large amount of experimental material is lost, it still results in a generic calibration model that is too cumbersome and leads to poor prediction results. At this time, finding an effective model upgrade and maintenance method can both reduce the sample consumption and make the model prediction accuracy improve, and also reduce the model complexity and make the maintained model more robust.
... These results suggested that the prediction of PLSR-SNV models should be considered in applications. According to Liu et al. (2015), a reasonable prediction should obtain an RPD value over 2. The proposed PLSR-SNV method achieved RPD >2 values, exceeding this recommended threshold, for all reference parameters. Hence, the PLSR-SNV models developed in this study for weight loss, HU and YI could be utilized for quality assessment, and most of the prediction models for the egg quality parameters could be considered T.T. Pham et al. reasonable. ...
... With sufficient training data, machine learning [18], [19], [20], [21] can project samples from spectral space to constituent concentration space. Additionally, if the predictions are to be made from indirect factors, such as gas constituents in a mixture being highly correlated to each other, machine learning will outperform conventional methods [15], [22]. In field applications such as on-site toxic gas detection and mineral analysis, memory size, processing speed and power consumption are limiting factors. ...
Through a study of multi-gas mixture datasets, we show that in multi-component spectral analysis, the number of functional or non-functional principal components required to retain the essential information is the same as the number of independent constituents in the mixture set. Due to the mutual in-dependency among different gas molecules, near one-to-one projection from the principal component to the mixture constituent can be established, leading to a significant simplification of spectral quantification. Further, with the knowledge of the molar extinction coefficients of each constituent, a complete principal component set can be extracted from the coefficients directly, and few to none training samples are required for the learning model. Compared to other approaches, the proposed methods provide fast and accurate spectral quantification solutions with a small memory size needed.
... Mastering the physiological characteristics of fruit structure and component distribution can obtain the best spectral measurement position and a more accurate prediction model (Pissard et al., 2012). Liu et al. (2015) found that the equatorial spectrum of Newhall navel orange was more suitable than the distal spectrum to establish a prediction model of fruit quality. Tian et al. (2020) found the best prediction position of SSC through regional combination based on differences in apple internal structure. ...
To quickly and accurately identify the quality of tomatoes, a method was proposed to predict the total soluble solid content (SSC), total titratable acidity (TA), and vitamin C (VC) content of tomatoes based on a multiregion combined model of the visible–near‐infrared spectrum. The results show that the competitive adaptive re‐weighted sampling algorithm combined with the partial least squares regression (CARS‐PLSR) model has the best prediction effect on SSC, TA, and VC content in “stem + equator”, “stem + bottom” and “stem + bottom” combinations. The prediction accuracy is 97.2%, 96.7%, and 97.7%, respectively, and the relative percent deviation (RPD) value is 5.870, 5.401, and 5.942, respectively.
Practical Application
This indicates that the CARS‐PLSR model based on the multiregion combination of visible–near‐infrared spectroscopy is reliable for predicting tomatoes' SSC, TA, and VC content. The results provide a theoretical basis for developing a portable fruit quality detector.
... The advantages of NIR spectroscopy are of great importance, nevertheless, there is the difficulty in analyzing and evaluating the NIR spectra, thus major efforts are being devoted to expanding our knowledge on multivariate analysis methods and their contribution to different NIR spectroscopy applications. A big part of these methods involves machine learning methods combined with preprocessing techniques [7], [8], [9], [10], [11]. ...
... Several applications of NIR combined with chemometrics are described in the literature for different fields, mainly food control [29][30][31]. NIR has been applied to predict sugar content in orange juice [32][33][34][35], passion fruit pulp [36,37], bayberry juice [38]. Regarding sugarcane, NIR has been used to predict the apparent sucrose content (soluble solids or polarimetric reading) using stalks [39,40] and juice [41][42][43]. ...
The aim of this work was to study dehydration as a way to improve the prediction of sucrose, glucose, and fructose in sugarcane juice using near-infrared (NIR) spectroscopy and partial least squares (PLS) regression models. The temperature, time, and sample volume involved in the dehydration process were optimized using design of experiments. Six different sample supports were assessed, being the thick couche paper the best support. NIR spectra from liquid (LSJ) and dehydrated sugarcane juice (DSJ) were obtained. Sucrose, glucose, and fructose in LSJ were analyzed using high-performance liquid chromatography with an evaporative light scattering detector (HPLC-ELSD). Sucrose, glucose, and fructose ranged from 99.29 to 249.27 mg/mL, 5.96–14.94 mg/mL and 3.99–16.10 mg/mL. PLS models were built using the sugars content and NIR spectra collected from a benchtop and a portable instrument. Ordered predictors selection (OPS) was applied to select the most informative variable. The results indicated better predictions for all sugars using the DSJ for both instruments, being the benchtop statistically better than the portable instrument. On the benchtop instrument, the PLS-OPS models presented root mean square error of prediction (RMSEP) respectively for sucrose, glucose, and fructose 7.98, 0.82, and 1.00 mg/mL using the DSJ against 12.75, 1.00, and 1.35 mg/mL using the LSJ. For the portable instrument, the RMSEP were respectively 15.90, 1.18, and 1.65 mg/mL using DSJ against 23.23, 1.40, and 2.08 mg/mL using LSJ. To sum up, the dehydration approach showed to be a great technique to improve the predictability of PLS-OPS models for sugarcane juice sugars using NIR spectra by removing the water and concentrating the analytes.
... Despite AsA's downward trend, a slight increase was observed in November in all varieties. NH had a higher AsA content than the other varieties throughout the sampling period, with statistically significant differences ( Figure 2A) according to previous reports [19,50]. ...
Sweet oranges are an important source of ascorbic acid (AsA). In this study, the content of AsA in the juice and leaves of four orange clonal selections, different in terms of maturity time and the presence/absence of anthocyanins, was correlated with the transcription levels of the main genes involved in the biosynthesis, recycling, and degradation pathways. Within each variety, differences in the above pathways and the AsA amount were found between the analysed tissues. Variations were also observed at different stages of fruit development and maturation. At the beginning of fruit development, AsA accumulation was attributable to the synergic action of l-galactose and Myo-inositol, while the l-gulose pathway was predominant between the end of fruit development and the beginning of ripening. In leaves, the l-galactose pathway appeared to play a major role in AsA accumulation, even though higher GalUr isoform expression suggests a synergistic contribution of both pathways in this tissue. In juice, the trend of the AsA content may be related to the decrease in the transcription levels of the GME, GDH, MyoOx, and GalUr12 genes. Newhall was the genotype that accumulated the most AsA. The difference between Newhall and the other varieties seems to be attributable to the GLDH, GalUr12, APX2, and DHAR3 genes.
... Several studies and publications have been carried out and reported regarding the application of NIRS method to determine some main quality parameters of foods and agricultural, meat and dairy products and others such as cocoa and chocolates [15][16][17][18], coffee [19][20][21], herbal plants [22][23][24], oranges [25,26], pineapple [27], milk and dairy products [28,29], cake [30], eggs [31], animal feed and meat [32,33]. ...
This study aims to apply near infrared technology as a fast, simultaneous and non-destructive method for quality assessment on intact mango fruit in form of total soluble solids (TSS) and vitamin C. Absorbance spectra of 186 intact mango fruits with four different cultivars were acquired and recorded in wavelength ranging from 1000–2500 nm. Spectra data were enhanced and corrected using three different methods namely moving average smoothing (MAS), extended multiplicative scatter correction (EMSC) and standard normal variate (SNV). In addition, they were divided into two datasets namely calibration (n = 143) and prediction (n = 43) datasets consisting all four mango cultivars. The models used to predict TSS and vitamin C were developed using partial least square regression (PLSR). Prediction performance were quantified using correlation coefficient (r), root mean square error (RMSE), ratio prediction to deviation (RPD) and range to error ratio (RER) indexes. The results showed that the best prediction models for TSS and vitamin C were achieved when the models were constructed using EMSC correction approach with r = 0.86, RMSE = 1.67 Brix, RPD = 2.34 and RER = 9.72 for TSS. Meanwhile, for vitamin C, r = 0.86, RMSE = 6.84 mg·100g⁻¹, RPD = 2.00 and RER = 8.87. From this study, it was concluded that near infrared technology combined with proper spectra enhancement method may be applied as a rapid, simultaneous and contactless method for quality assessment on intact mangoes. © 2021. the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0). All Rights Reserved.
... Various feature extraction algorithms [11] followed by different chemo-metric algorithms [12] have been used to extract important information and to correlate with various quality parameters. Other than chemo-metric algorithms, researchers also have been introduced different machine learning and deep learning-based modelling schemes [13] to correlate features of the processed spectral dataset with various quality parameters such as chlorophyll, acidity, total soluble solid, etc. ...
Optimum maturity and ripeness at the time of harvest is highly important to maintain the nutritional parameters of fruits. Maturity and ripeness of most of the fruit samples depends on various physiochemical parameters such as color, shape, size, total soluble solid and many more. Several state-of-the-art solutions such as GC–MS, Electronic Nose, Spectrometer and many more are available to measure various fruit quality parameters but most of the solutions available in the market are bulky, time consuming, lab-level and requires skilled manpower for operation. Presented manuscript reports a battery operated, smartphone spectrometer based solution to carry out the variety of activities in the field. Overall device uses UV–Vis-NIR led array as source and collection of spectral sensors (AS7262 and OPT101) to acquire overall UV–Vis-NIR spectrum over the range of 400–1000 nm with the resolution of 40 nm. Designed source and detector modules have been interfaced with designed triggering, filter and amplification circuit. A low power wireless solution along with on-board microcontroller facility has been designed and interfaced with circuits, source and detectors. All essential components such as source, detectors, filters, lens and all circuits have been assembled in a housing of dimensions 18.0 × 9.0 × 6.0 (in cm) and the entire device weighs 183.35 g. Different statistical and neural network based modelling techniques have been explored to design prediction models for total soluble solids, weight, volume, chlorophyll, sugar content and acidity. Models have been evaluated based on accuracy, memory and time usage. Best performed models have been used to train handheld smartphone based spectrometer device to predict various quality parameters for citrus samples. System communicates data to smartphone based android app to display various parameters. Android app also provides facility to save data on cloud with tree and orchard ID to monitor overall yield and harvesting time.
... Cayuela and Weiland (2010), using two different portable NIR-spectrometers, built SSC and acidity prediction models in a non-destructive way, reaching R 2 = 0.91 and 0.83, respectively. Liu et al. (2015), in a study, using NIR to determine SSC, pH, TA (titratable acidity), and vitamin C in Newhall navel oranges, developed satisfactory prediction models, by LS-SVM (Least Squares Support Vector Machine) with a prediction correlation coefficient higher than 0.82. NIR has also been used to quantify vitamin C contents in oranges, showed low RMSEP with RMSEV = 3.90 mg 100 g -1 (Jun-fang et al. 2007). ...
Near (NIR) and mid (MIR) infrared spectroscopies have been studied as potential methods for non-destructive analyses of the fresh fruits quality. In this study, vitamin C, citric acid, total and reducing sugar content in ‘Valência’ oranges were evaluated using NIR and MIR spectroscopy with multivariate analysis. The spectral data were used to build up prediction models based on PLS (Partial Least Squares) regression. For vitamin C and citric acid, both NIR (r = 0.72 and 0.77, respectively) and MIR (0.81 and 0.91, respectively) resulted in feasible models. For sugars determination the two techniques presented a strong correlation between the reference values and analytical signals, with low RMSEP and r > 0.70 (NIR: sucrose RMSEP = 12.2 and r = 0.75; glucose RMSEP = 6.77 and r = 0.82; fructose RMSEP = 5.07 and r = 0.81; total sugar RMSEP = 12.1 and r = 0.80; reducing sugar RMSEP = 20.32 and r = 0.82; MIR: sucrose RMSEP = 9.47 and r = 0.80; glucose RMSEP = 6.70 and r = 0.82; fructose RMSEP = 5.20 and r = 0.81; total sugar RMSEP = 11.72 and r = 0.81; reducing sugar RMSEP = 20.42 and r = 0.81). The models developed with MIR presented lower prediction error rates than those made with NIR. Therefore, infrared techniques show applicability to determine of orange quality parameters in a non-destructive way.
... It could be seen that the Smoothing had no significant difference with original spectra because their latent variables (LVs), R pre , RMSEP and RPD values were close, indicating that Smoothing unable to remove Journal Pre-proof J o u r n a l P r e -p r o o f undesirable signal interference in original spectra. The first derivative is not applicable to preprocess the original transmittance spectra of citrus, because the performance turned worse than the original spectra, the result was similar to Liu et al. (2010) and Liu et al. (2015). The best performance of the prediction model was derived from the pretreated spectra of SNV with R pre of 0.9215, RMSEP of 0.5546 °Brix, and RPD of 2.5746, respectively. ...
Nondestructive determination the internal quality of thick-skin fruits has always been a challenge. In order to investigate the prediction ability of full transmittance mode on the soluble solid content (SSC) in thick-skin fruits, the full transmittance spectra of citrus were collected using a visible/near infrared (Vis/NIR) portable spectrograph (550–1100 nm). Three obvious absorption peaks were found at 710, 810 and 915 nm in the original spectra curve. Four spectral preprocessing methods including Smoothing, multiplicative scatter correction (MSC), standard normal variate (SNV) and first derivative were employed to improve the quality of the original spectra. Subsequently, the effective wavelengths of SSC were selected from the original and pretreated spectra with the algorithms of successive projections algorithm (SPA), competitive adaptive reweighted sampling (CARS) and genetic algorithm (GA). Finally, the prediction models of SSC were established based on the full wavelengths and effective wavelengths. Results showed that SPA performed the best performance on eliminating the useless information variable and optimizing the number of effective variables. The optimal prediction model was established based on 10 characteristic variables selected from the spectra pretreated by SNV with the algorithm of SPA, with the correlation coefficient, root mean square error, and residual predictive deviation for prediction set being 0.9165, 0.5684°Brix and 2.5120, respectively. Overall, the full transmittance mode was feasible to predict the internal quality of thick-skin fruits, like citrus. Additionally, the combination of spectral preprocessing with a variable selection algorithm was effective for developing the reliable prediction model. The conclusions of this study also provide an alternative method for fast and real-time detection of the internal quality of thick-skin fruits using Vis/NIR spectroscopy.
... In the original IRC model, the solution to the solution vector x i usually uses the least squares fitting method, and the optimization target usually reconstructs the error under the L 2 norm ‖ ‖ − i i ‖ ‖2 and make it as small as possible [24]. The L 2 norm, that is, the Euclidean distance, can overcome the influence of partial noise on the classification of samples, but often ignores the statistical distribution characteristics inside the sample. ...
In order to establish an accurate and efficient model for geographical origin identification of oranges, a new model based on L1-norm linear regression classification (L1-LRC) is proposed. The proposed L1-LRC for orange origin identification is based on minimum reconstruction error using the L1-norm regularization learning method, which can combine the feature selection and classifier learning, and can reveal the structure characteristics of spectral information effectively. The experimental results show that the proposed L1-LRC model can achieve higher accuracy rate of 92.35% and perform much better than existing models when using only a few training samples. Thus, this work would lead to a new method for fast and efficient identification of geographical origins with near infrared (NIR) spectroscopy.
... Many classification and machine learning methods have been used for measurement data analyses. Recent studies include machine learning with optics, NIR and X-rays [27][28][29][30][31][32]. Electromagnetic waves using radio and microwave frequencies have been used with machine learning based classification techniques [33,34]. ...
Wood chips are extensively utilised as raw material for the pulp and bio-fuel industry, and advanced material analyses may improve the processes in utilizing these products. Electrical impedance spectroscopy (EIS) combined with machine learning was used in order to analyse heartwood content of pine chips and bark content of birch chips. A novel electrode system integrated in a sampling container was developed for the testing using frequency range 42 Hz-5 MHz. Three electrode pairs were used to measure the samples in x-, y-and z-direction. Three machine learning methods were used: K-nearest neighbor (KNN), decision tree (DT) and support vector machines (SVM). The heartwood content of pine chips and bark content of birch chips were classified with an accuracy of 91% using EIS from pure materials combined with a k-nearest neighbour classifier. When using mixed materials and multiple classes, 73% correct classification for pine heartwood content (four groups) and 64% for birch bark content (five groups) were achieved.
... The results of the LS-SVM model presented an accurate estimation of ascorbic acid in apple (R 2 p = 0.80). A similar result of LS-SVM (R p = 0.83) was observed when NIR was used for the prediction of ascorbic in orange samples (Liu et al. 2015b). Additionally, this technique was applied in the determination of ascorbic acid in summer squash with R 2 p of 0.86 ) and passion fruit with R p of 0.663 (Maniwara et al. 2014). ...
Nowadays, progresses in data processing software have promoted the application of infrared (e.g., FT-IR, NIR, MIR), Raman, and hyperspectral imaging (HSI) techniques for quantitative analysis of biological material and/or aroma compounds in foods. In this review, applications of vibrational spectroscopy combined with chemometrics are summarized including analysis of total polyphenol, individual polyphenols, vitamins, and aromatic compounds in raw and some processed products. Laboratory-based and online application of vibrational spectroscopies monitoring for analysis of phenolic compounds have been described. In addition, technical challenges and future trends have been covered. Based on the literature, the near-infrared technique often has an advantage over other spectroscopy approaches and the expensive and time-consuming chemical methods such as high-performance liquid chromatography and gas chromatography. Overall, the current review suggests that vibrational spectroscopies are promising and powerful techniques that can be used for rapid and accurate determinations of food nutraceuticals and volatile compounds in both academic and industrial contexts.
... NIR is highly penetrative and no sample preparation is needed before the analysis (Srivastava and Sadistap, 2018). Intense research has been carried out with NIR spectroscopy for maturity classification depending on the given quality parameters of tropical fruits such as acerola, jujube, kiwi, mango, passion fruit, pomelo, or fruits and vegetables that have high economic value due to mass consumption rates such as apple, carrot, cucumber, grapes, orange, pumpkin, strawberries, tomatoes, etc. (Jamshidi et al., 2012;Sánchez et al., 2012;Sirisomboon et al., 2012;Jha et al., 2014;Liu et al., 2015;Chen et al., 2016;Guo et al., 2016;Rungpichayapichet et al., 2016;Malegori et al., 2017;Ncama et al., 2017;Beghi et al., 2018;Cavaco et al., 2018;Oliveira-Folador et al., 2018;Xiao et al., 2018). ...
In this chapter, non-destructive spectroscopic, computed tomography, mechanical, and other miscellaneous methods were discussed based on fruit and vegetable applications.
The implementation cost of equipment, the length of data processing, as well as the amount of data to be extracted, push ahead the technological improvements to build portable, practical, on-line, real-time, in-field, and inexpensive methods. Precise, accurate, and repeatable measurements are required for correct classification of foods; however, in most cases the amount of data is voluminous, and employing appropriate preprocessing methods and algorithms may increase the accuracy of results and decrease the complexity of data to be processed. Several recent non-destructive applications have been discussed, and each of them has some limitations or advantages over other methods. The continuous developments in equipment and software technology, especially introduction of artificial intelligence, for in-line measurements provides fast and accurate prediction of food quality, thus avoidable food losses will gradually be decreased the with these improvements.
... Oleh sebab itu, diperlukan suatu metode yang dapat mengetahui kandungan ALB secara kuantitatif. NIR spectroscopy merupakan salah satu metode non-destruktif yang dapat memprediksi kualitas internal bahan pertanian secara cepat dan tepat (Cayuela JA, 2017;Li M, 2017;Liu C, 2015). Metode ini juga dilaporkan dapat mengetahui perubahan tingkat kematangan buah dengan memprediksi kandungan kimia pada tahap kematangan (Jha SN, 2014;Watanawan C, 2014). ...
Kadar Asam Lemak Bebas (ALB) yang rendah merupakan salah satu indikator kualitas Crude Palm Oil (CPO) yang baik. Apabila Tandan Buah Segar (TBS) kelapa sawit yang lewat matang ikut diolah menjadi CPO, maka kadar ALB selama produksi dapat meningkat. Proses pemanenan menjadi titik krusial yang sangat mempengaruhi tingkat kematangan buah. Selama ini penentuan kematangan TBS kelapa sawit masih dilakukan secara visual yang bergantung kepada kemampuan dan kondisi pemanen buah sawit. Oleh karena itu, perlu dikembangkan suatu metode secara kuantitatif yang dapat memprediksi kadar ALB secara objektif. Pada penelitian ini, akan dikembangankan metode non-destruktif berbasis NIR spectroscopy yang akan dikaji sebagai metode untuk menentukan tingkat kematangan TBS berdasarkan kandungan ALB. Penelitian ini dibagi menjadi tiga tahapan, yaitu akuisisi data reflektansi spektrum TBS dengan NIR Flex N-500, pengukuran kadar ALB, dan pembangunan model kalibrasi dengan menggunakan kemometrik. Dari hasil pengembangan model didapatkan nilai R2 tanpa preprocessing sebesar 0.236, RPD sebesar 1.27 dan PC sebesar 2. Proprocessing First Derivative Savitzky Golay (DG1) memberikan nilai koefisien determinasi tertinggi yaitu sebesar 0.243, dengan nilai RPD sebesar 1.17 dan PC sebesar 2. Akan tetapi kualitas model kalibrasi yang dibangun tetap belum mampu menunjukkan kehandalan dalam memprediksi kandungan ALB tandan buah segar kelapa sawit.
... As regards ascorbic acid content, no references have been found where this parameter has been measured in spinach using NIRS. However, some authors [21][22][23][24][25] have shown how NIRS technology can be used to measure ascorbic acid in apples, zucchini, oranges, potatoes and peppers. ...
The study sought to perform a non-destructive and in-situ quality evaluation of spinach plants using near infrared (NIR) spectroscopy in order to establish its suitability for different uses once harvested. Modified partial least square (MPLS) regression models using NIR spectra of intact spinach leaves were developed for nitrate, ascorbic acid and soluble solid contents. The residual predictive deviation (RPD) values were 1.29, 1.21 and 2.54 for nitrate, ascorbic acid and soluble solid contents, respectively. Later, this predictive capacity increased for nitrate content (RPDcv = 1.63) when new models were developed, taking into account the influence on the robustness of the model exercised by the simultaneity between the NIR and laboratory analyses. Subsequently, using partial least squares discriminant analysis (PLS-DA), the ability of NIRS technology to classify spinach as a function of nitrate content was tested. PLS-DA yielded percentages of correctly classified samples ranging from 73.08-76.92% for the class 'spinach able to be used fresh' to 85.71-73.08% for the class 'preserved, deep-frozen or frozen spinach, both for unbalanced and balanced models respectively, based on NH signal associated with proteins. Overall, the data supports the capability of NIR spectroscopy to establish the final destination of the production of spinach analysed on the plant, as a screening tool for important safety and quality parameters.
... Machine learning is a research field related to the automatic pattern recognition from a given knowledge database towards making decisions and predictions. 21 Some recent researches in the agricultural field have used NIR and ML tools with good prediction results becoming a new alternative for the traditional form of validation. 22 Therefore, the present work aims at contributing in the development of a rapid and nondestructive approach to classify different types of chicken parts, which can be potentially used in chicken products adulteration, through the use of NIR technique performed by a portable spectrophotometer associated with ML. ...
Identification of different chicken parts using portable equipment could provide useful information for the processing industry and also for authentication purposes. Traditionally, physical–chemical analysis could deal with this task, but some disadvantages arise such as time constraints and requirements of chemicals. Recently, near-infrared (NIR) spectroscopy and machine learning (ML) techniques have been widely used to obtain a rapid, noninvasive, and precise characterization of biological samples. This study aims at classifying chicken parts (breasts, thighs, and drumstick) using portable NIR equipment combined with ML algorithms. Physical and chemical attributes (pH and L*a*b* color features) and chemical composition (protein, fat, moisture, and ash) were determined for each sample. Spectral information was acquired using a portable NIR spectrophotometer within the range 900–1700 nm and principal component analysis was used as screening approach. Support vector machine and random forest algorithms were compared for chicken meat classification. Results confirmed the possibility of differentiating breast samples from thighs and drumstick with 98.8% accuracy. The results showed the potential of using a NIR portable spectrophotometer combined with a ML approach for differentiation of chicken parts in the processing industry.
... Least angle regression (LAR) is an algorithm proposed by Efon et al. in 2004 to achieve the feature selection and linear regression [17], which is rarely used in the field of spectral analysis. Liu et al. [18] performed a regression prediction experiment using the LAR method on the navel orange near infrared data. The results showed that LAR is better than PLS in prediction performance and interpretation, and is faster than least squares support vector machine (LS-SVM) at the running speed. ...
In this paper, a regression analysis method based on the combination of Least Angle Regression (LAR) and Partial Least Squares (PLS) is proposed, which uses the non-invasive characteristics of near infrared spectroscopy (NIRS) to implement early screening of leukemia patients. First, the LAR method is used to eliminate collinearity between variables, second, PLS is employed to further build model for the wavelengths which are selected by the LAR. The result shows that this method needs less wavelength points and has more excellent performance in correlation coefficient and root mean square error, that are 0.9492 and 0.5917 respectively. The comparison experiments demonstrate that the LAR-PLS regression model has an advantage over principal component regression (PCR), the LAR-PCR regression model, successive projections algorithm (SPA) and elimination of uninformative variables (UVE) combined with PLS method in terms of predictive accuracy for screening leukemia patients.
... In the last few years, near infrared (NIR) spectroscopy has gained wide acceptance in many research fields as a simple, quick and non-destructive technique that allows the identification of chemical compounds from samples without previous preparation (Craig et al., 2014;Liu et al., 2015;Zhang et al., 2013). NIR has been coupled with multivariate techniques aimed to identify patterns emerging from different products with categorization purposes in several fields, including food (Karoui and De Baerdemaeker, 2007;Marcelo et al., 2014b), pharmaceuticals (Anzanello et al., 2013;Gendrin et al., 2007) and fuels (de Vasconcelos et al., 2012;Ferrão et al., 2011). ...
Yerba mate (Ilex paraguariensis) is used to produce a beverage typically consumed in South America countries, and presents peculiar land-based characteristics due to geographical origin. Such characteristics have recently become a matter of interest for many producers as specific features of yerba mate tend to influence product acceptance in new markets, prices and commercial advantages. This scenario justifies the developing of frameworks tailored to correctly classify products according to their authenticity. This paper uses Near Infrared (NIR) spectroscopy and data describing concentration of chemical elements to classify commercial yerba mate samples according to their place of origin. Aimed at enhancing data interpretability, we propose a novel variable selection method that applies quadratic programming to reduce redundant information among the retained variables and maximize their relationship regarding the sample place of origin; sample categorization is then performed using alternative classification techniques. When applied to the NIR dataset, the proposed method retained average 8.79% of the original wavenumbers, while leading to 1.9% more accurate classifications when compared to categorization using the full spectra. As for the elements dataset, we increased average classification accuracy by 3.5% and retained 47.22% of the original elements. The proposed method also outperformed two other approaches for variable selection from the literature. Our findings suggest that variable selection frameworks help to correctly identify the origin and authenticity of yerba mate samples, making model construction and interpretation easier.
... Since intact fruit were scanned, the deduction made was that there was high noise effect in the 450-800 nm wavelength range. This is probably due to thick citrus peel interfering with spectral information for internal quality parameters and attributing to noise in the visible range (Magwaza et al., 2013;Wang et al., 2014;Liu et al., 2015). The low predictability of models developed using the visible range may also suggest limited penetration depth in this region of the spectrum (Lammertyn et al., 2000). ...
Sweetness and flavour are desirable attributes used for quality control and assurance of citrus fruit, which are largely determined by total soluble solids (TSS), titrable acidity (TA) and TSS: TA ratio. However, the accuracies of TSS, TA and TSS: TA as flavour indices have been recently criticised. BrimA (Brix minus acids), on the other hand, is an accurate organoleptic parameter that has been shown to be highly related to sweetness and flavour of citrus fruit. In this study, the ability of visible to near infrared spectroscopy (Vis/NIRS), in reflectance mode, to non-destructively quantify BrimA, TSS, TA and TSS: TA ratio of 'Valencia' orange and 'Star Ruby' grapefruit was evaluated. Vis/NIR spectral data was acquired using a laboratory bench-top monochromator NIR Systems. Reference measurements and spectral datasets were subjected to partial least square (PLS) regression analysis. The best prediction model were observed for BrimA of 'Va-lencia' oranges with the coefficient of determination (R 2) = 0.958; root mean square error of prediction (RMSEP) = 0.006 and residual predictive deviation (RPD) = 3.96, following TSS: TA ratio (R 2 = 0.958; RMSEP = 0.605; RPD = 4.92). Good models for predicting flavor of grapefruit were also attained, with TSS having the best model (R 2 = 0.896, RM-SEP = 0.308 and RPD = 2.94), followed by BrimA (R 2 = 0.858; RMSEP = 0.429; RPD = 2.45). These results demonstrated the ability of Vis/NIRS to non-destructively predict sweetness and flavour attributes of oranges and grapefruit. Vis/NIRS was recommended as a possible fast and accurate technique to be used for fruit discrimination based on flavour parameters during packing and for pricing of fruit in the market.
... The classification of food samples based on their chemical composition provides useful information for a variety of purposes, such as recognition of geographical origin and authenticity, the characteristics of a product, quality control for companies, preservation, and category differentiation (Barbosa et al., 2015;Barbosa, Nacano, Freitas, Batista, & Barbosa, 2014b;Khanmohammadi et al., 2014;Liu, Yang, & Deng, 2015;Maione, Batista, Campiglia, Barbosa, & Barbosa, 2016;Qiu, Wang, Tang, & Du, 2015;Soltani & Omid, 2015). When it comes to organic food, authenticity is a substantial matter and a current concern of the organic industry. ...
Non-destructive identification of the fungal disease has great importance as one of the challenges of the citrus post-harvest industry. In order to identify the fungal disease of Alternaria alternata, which occurred internally, the reflectance spectra were collected in 3 areas of the stem-end, equatorial and stylar-end of healthy and fungus-infected oranges using VIS-NIR spectrometer (400–1100 nm). Principal Component Analysis models were extracted from the preprocessed wavelengths after performing various preprocessors such as Savitzky-Gulay, standard normal variate and mean normalization. The classification capability of support vector machine and back-propagation neural network classifiers was evaluated in identifying appropriate preprocessors from the obtained principal components and identifying infected fruits using optimal wavelengths. The highest accuracy of the classification was obtained by the back-propagation neural network classifier using the optimal wavelengths of the stylar-end area. It included training, validation and testing stages for Thompson cultivar that was 94%, 90% and 93% and for Jaffa cultivar, was 97%, 91% and 97%, respectively. Appropriate classification results indicate the generalizability of the models developed to identify internal fungal infection of Alternaria alternata in orange fruit.
Background and Objectives
Grain quality is a complex trait in rice, compared with other staple crops as it is predominantly consumed as a whole grain. Although considered secondary to yield, to align with consumer preferences, breeders are increasingly interested in quality. At the early stages of a breeding program, grain quality‐related traits are often ignored as they are arduous and time‐consuming. Near‐infrared spectroscopy (NIRS) could be a suitable high‐throughput alternative to conventional wet chemistry and image processing‐related methods to be adopted for early screening. This study aims to quantify traits essential for rice breeders such as amylose, chalkiness, length, width, and the length/width ratio in rice samples with NIRS. We used conventional algorithms such as principal component analysis (PCA), partial least square regression (PLSR), multilayer perceptron (MLP), support vector classification (SVC), and linear discriminant analysis (LDA) to compare with the proposed convolutional neural network (CNN) for regression and classification.
Findings
Our results showed that the proposed CNN outperformed the conventional models in estimating all traits. Unlike conventional models, CNN models could be developed with raw spectra with minimal to no preprocessing, and along with the transfer‐learning capabilities, the time required for model development could be significantly reduced.
Conclusion
We recommend NIRS for quantitative estimation of amylose and chalkiness in rice and rather use classification/categorized estimation for other physical dimension‐related traits such as length and length/width ratio.
Significance and Novelty
We found NIRS to be an appropriate alternative to wet chemistry and image‐based methods for screening lines at the early stages of the breeding program. Estimation of physical parameters such as length and length/width ratio with NIRS is novel and appears reasonable for high‐throughput applications.
The increasing awareness regarding dietary diversity and pattern has increased the demand for quality over quantity. Numerous non-destructive measurements, including visible near-infrared spectroscopy, have been used in assessing the soluble solid content (SSC) in foods. With advances in statistics, various statistical methods have been developed. These methods need to be verified for their application in effective SSC prediction models. This study aims to review the utility of various statistical methods for the SSC prediction of apples. In this study, we constructed a sorting device for Fuji apples. The spectra of the apples were measured, and the potential of the SSC prediction model was evaluated using various pre-processing methods and machine learning techniques. A developed support vector regression model with a first-order derivative method exhibited the highest prediction accuracy (R² = 0.8503, RMSEP = 0.4781). Therefore, the developed efficient spectrum pre-processing method coupled with a robust machine learning model was useful for improving the prediction performance of the sorting device.
Apple, as an important agricultural product, has extremely high nutritional value. In order to distinguish apple varieties quickly, accurately, and nondestructively, an improved possibilistic Gath–Geva (IPGG) clustering algorithm was proposed to classify near infrared reflectance (NIR) spectra of apple samples. This paper used Antaris II NIR spectrometer (Thermo Electron Co., USA) to collect NIR spectra of four kinds of apples (Fuji, Huaniu, Gala, and Huangjiao). Then, multiple scatter correction (MSC) and principal component analysis (PCA) were applied to eliminate redundant information and reduce spectral dimensions, respectively. Finally, fuzzy c‐means (FCM), Gustafson‐Kessel (GK), Gath–Geva (GG), improved possibilistic c‐means (IPCM), and IPGG clustering algorithms were run on the preprocessed spectral data. The results shown that the clustering accuracy of IPGG was the highest, and it reached 96.5%. Experimental results demonstrated that NIR spectroscopy along with MSC, PCA, and IPGG clustering was an effective method for identifying apple varieties. The apple variety is of great importance to the quality of apple. For this, the proposed IPGG clustering along with near infrared reflectance spectroscopy was used to build an effective classification model to identify apple varieties quickly, accurately, and nondestructively. The experimental results showed that IPGG clustering algorithm has obvious advantages compared with FCM, GK, GG, and IPCM. This study provides a new method for apple grading and screening at the fruit and vegetable processing plants.
In Near-infrared (NIR) spectroscopy qualitative analysis, noise caused data quality problem has been a bottleneck to further enhance the prediction accuracy. The appropriate preprocessing methods can reduce the influence of noise; and robust models have higher tolerance for noise disturbance. But these methods treat all the wavelengths equally. In fact, the spectra at different wavelengths may have very different level of noise. In this paper, it is proposed a noise level penalizing robust Gaussian process (NLP-RGP) regression on NIR spectroscopy quantitative analysis. The noise level penalizing mechanism penalize the spectra features according to their noise level, i.e., encourage the model to prefer the less noisy features over high noisy features. Gaussian process (GP) is a nonparametric machine learning method based on kernel and Bayesian inference framework; with a noise model of heavy-tailed distribution, robust Gaussian process can handle the abnormal sample data better. Experiments were taken on the determination of the total soluble solids content of navel oranges based on their surface NIR spectra. Robust Gaussian process (RGP) performs better than common Gaussian process model, and noise level penalizing Gaussian process (NLP-GP) performs even better than the robust Gaussian process model. Furthermore, NLP-GP outperforms least squares support vector machines (LS-SVM), the state of art method. Among all the models, the NLP-RGP achieves the best prediction accuracy.
Near-infrared (NIR) spectroscopy as a low-cost technique with its non-destructive fast nature, precision, control, accuracy, repeatability, and reproducibility has been extensively employed in most industries for food quality measurements. Its coupling to different modeling techniques has been identified as a way of improving the accuracy and robustness of non-destructive measurement of foodstuffs. This review provides an overview of the application of non-linear algorithms in food quality and safety specific to NIR spectroscopy. The review also provides in-depth knowledge about the principle of NIR spectroscopy along with different non-linear models such as artificial neural network (ANN), AdaBoost, local algorithm (LA), support vector machine (SVM), and extreme learning machine (ELM). Moreover, non-linear algorithms coupled with NIR spectroscopy for ensuring food quality and their future perspective has been discussed.
Rapid determination of moisture content during processing period of apple chips is of great significance for their real-time quality control. In this study, near infrared spectroscopy (NIRS) combined with chemometric techniques were used to measure moisture content of apple chips during processing period. Interval partial least squares (iPLS), synergy interval partial least squares (siPLS) and genetic algorithm (GA) were utilized to select efficient wavelength regions, respectively, and partial least squares regression (PLSR) was applied to develop prediction model. The performance of models was evaluated by three types of parameters, including the root mean square error (RMSE), correlation coefficient (R) and the residual prediction deviation (RPD). The results demonstrated that, in contrast to classical PLS, iPLS and GA-PLS model, siPLS model for the prediction of moisture content had superior performance with RMSEC = 0.0877, Rc = 0.9540, RMSEP = 0.0627, Rp = 0.9747, and RPD = 4.37. Therefore, the application of NIRS and wavelength selection techniques in this study provides a feasibility of a non-destructive and rapid way to enhance the efficiency of quality assurance and control investment.
Quality control (QC) of food includes control of authenticity, origin, properties or attributes, composition, and safety involving all steps of the global supply chain – from the raw material up to the final product sold on the supermarket shelf. Despite the great concern regarding their quality, these products are currently monitored and certified by means of paper documentation, some targeted analysis, and sensory properties. This article encompasses a revision on the instrumental analytical techniques, methods, and systems used in food QC together with their main applications and a brief historical perspective on food analysis, together with a deep revision on the current state of art of modern analytical instruments, methodologies, and applications in the QC of food in order to establish Quo Vadis QC of foods. This article also discusses the present and future challenges in QC of food, the need for faster techniques, the prospect of advanced imaging techniques, and the application of ‘omic’ in food QC (including metabolomics, proteomics, etc.)
The classification of growing locations is very important for quality control in the orange industries, which is also challenging work, because of its complex chemical composition and varies of taste and sizes. The traditional ways to classify them by human’s sense are time consuming and at high cost. In this paper, a new general classification framework based on the Near-Infrared Reflection (NIR) spectroscopy using data mining technology was proposed. First, the raw NIR spectra data were reduced by the principal components analysis (PCA), and then an attribution selection method was applied to find the best feature subset. An evolution process was also introduced to test the performance of five classifiers (Decision Tree, KNN, Naive Bayesian, SVM and ANN) used in this paper. The proposed classification framework was verified on three NIR spectra datasets, which were collected from the different part of oranges (including two parts of fruit surface and juice) from 15 different places in china. The experimental results demonstrated that the juice NIR spectra is the most suitable data-set for identifying the orange growing locations, and the decision tree is the best and most stable classifier, which could achieve the highest average prediction rate of 96.66%.
This study proposes a formula for prediction of maturity index (Im) using physico-chemical characteristics and overall acceptability (OA) of a sensory panel for mangoes from orchards of nine Indian states. Computed Im values were found to be in agreement with both OA scores and the perceptions of experienced farmers. NIR spectra of 1180 mangoes were acquired. Multiple-linear regression (MLR) and partial least square (PLS) models were developed in the wavelength range of 1200–2200 nm to predict Im. The best prediction was achieved using PLS model after MSC data treatment in the wavelength range of 1600–1800 nm. Multiple correlation coefficients (R) for calibration and validation of PLS model were 0.74 and 0.68, respectively. Lower difference in standard errors of calibration (0.305) and prediction (0.335), indicated the potential of NIRS in prediction of the maturity non-destructive.
External and internal quality parameters were measured in oranges (Citrus sinensis (L.) Osbeck cv. “Powell Summer Navel”) during on-tree ripening and at harvest using near-infrared reflectance (NIR) spectroscopy. The performance of two NIRS instruments was evaluated: a handheld microelectromechanical system spectrophotometer working in the 1,600- to 2,400-nm range, and a diode array visible–NIR spectrophotometer working in the 380- to 1,700-nm range. Spectra and analytical data were used to construct MPLS prediction models for quantifying weight, size (equatorial and axial diameters), color (L*, a*, b*, C*, h*, and color index), texture (firmness and maximum penetration force), yield (pericarp thickness, juice weight, and juice content), and chemical parameters (soluble solids content, pH, titratable acidity, and maturity index). Both instruments yielded promising results for on-tree and at-harvest quality measurements, but models constructed using the diode array instrument provided greater predictive capacity, particularly for fruit size (equatorial and axial diameters) and total soluble solids content. Subsequent evaluation of the LOCAL algorithm revealed that it increased the predictive capacity of models constructed for all the main parameters tested. These results confirm that noninvasive NIRS technology can be used to simultaneously evaluate external and internal quality parameters in intact oranges both during on-tree ripening and at harvest, thus making it easier for farmers to monitor the ripening process and also to optimize harvest timing in order to meet the demands of the citrus-fruit industry.
The global citrus industry is continually confronted by new technological challenges to meet the ever-increasing consumer
awareness and demand for quality-assured fruit. To face these challenges, recent trend in agribusiness is declining reliance
on subjective assessment of quality and increasing adoption of objective, quantitative and non-destructive techniques of quality
assessment. Non-destructive instrument-based methods are preferred to destructive techniques because they allow the measurement
and analysis of individual fruit, reduce waste and permit repeated measures on the same item over time. A wide range of objective
instruments for sensing and measuring the quality attributes of fresh produce have been reported. Among non-destructive quality
assessment techniques, near-infrared (NIR) spectroscopy (NIRS) is arguably the most advanced with regard to instrumentation,
applications, accessories and chemometric software packages. This paper reviews research progress on NIRS applications in
internal and external quality measurement of citrus fruit, including the selection of NIR characteristics for spectra capture,
analysis and interpretation. A brief overview on the fundamental theory, history, chemometrics of NIRS including spectral
pre-processing methods, model calibration, validation and robustness is included. Finally, future prospects for NIRS-based
imaging systems such as multispectral and hyperspectral imaging as well as optical coherence tomography as potential non-destructive
techniques for citrus quality assessment are explored.
KeywordsNon-destructive evaluation–Near-infrared spectroscopy–NIRS–Citrus fruit–Internal quality–External quality–Hyperspectral–Multispectral–Optical coherence tomography (OCT)–X-ray computed tomography (CT)
Sugar content is one of the most important quality attributes of citrus fruit, either for fresh or for processing market.
Since sugars in citrus juice are highly correlated with total soluble solids (TSS) content, which can be determined easily
even by the means of a hand refractometer, TSS is one of the most frequently used quality index. Since TSS can be measured
only destructively, the results are representative only if carried out on large samples and do not allow classifying marketable
fruit one by one according to their specific sugar content. Objective of this experiment was to assess possibility and limits
of a non-destructive estimation of citrus fruits internal quality parameters (TSS and titratable acidity) presenting thick
peel by the use of a spectrophotometric portable VIS-NIR system. Four hundred fruit of “Miho” satsuma and 150 fruit of “Page”
tangelo were used. Each fruit was first subjected to spectrophotometric acquisition and soon after was juiced and TSS and
titratable acidity (TA) determined. Partial least squares (PLS) regression analysis was applied for constructing a predictive
model based on the spectral normalized response, constructing the model on a sub-sample and verifying the model (prediction
test) on independent ones. The TA relative to Page mandarin was predicted in the test with an r = 0.88 and a standard error of prevision (SEP) coefficient of variability of 3.8% while the TSS scored an r = 0.85 and a SEP coefficient of variability equal to 4%. The TA of Miho mandarin was predicted in the test with an r = 0.81 and a SEP coefficient of Variability of 8.3% while the TSS scored an r = 0.84 and a SEP coefficient of variability equal to 5.6%.
KeywordsMandarin–VIS-NIR–Partial least squares–Total soluble solids–Titratable acidity
Support Vector Machines Basic Methods of Least Squares Support Vector Machines Bayesian Inference for LS-SVM Models Robustness Large Scale Problems LS-SVM for Unsupervised Learning LS-SVM for Recurrent Networks and Control.
A portable near infrared spectroscopy system was developed for assessing the quality of Nanfeng mandarin fruit. One hundred and fifty-three Nanfeng mandarin samples were used to measure the performance of the system. Several pretreatment methods were adopted to process the spectra. Then Support Vector Machine (SVM), Back Propagation Neural Network (BPNN) and Partial Least Square (PLS) were used to build models for soluble solids content (SSC), titratable acidity (TA), vitamin C and surface color. The best results were obtained by SVM. The correlation coefficient (R) and root mean square error of prediction (RMSEP) were (0.93, 0.65°Brix), (0.66, 0.09%), (0.81, 2.7mg/100g) and (0.57, 0.81) for SSC, TA, vitamin C and color, respectively. The results demonstrated that the portable near infrared spectroscopy was feasible for determining the Nanfeng mandarin quality nondestructively.
Breast cancer is categorized into two broad groups: estrogen receptor positive (ER+) and ER negative (ER-) groups. Previous study proposed that under trastuzumab-based neoadjuvant chemotherapy, tumor initiating cell (TIC) featured ER- tumors response better than ER+ tumors. Exploration of the molecular difference of these two groups may help developing new therapeutic strategies, especially for ER- patients.
With gene expression profile from the Gene Expression Omnibus (GEO) database, we performed partial least squares (PLS) based analysis, which is more sensitive than common variance/regression analysis.
We acquired 512 differentially expressed genes. Four pathways were found to be enriched with differentially expressed genes, involving immune system, metabolism and genetic information processing process. Network analysis identified five hub genes with degrees higher than 10, including APP, ESR1, SMAD3, HDAC2, and PRKAA1.
Our findings provide new understanding for the molecular difference between TIC featured ER- and ER+ breast tumors with the hope offer supports for therapeutic studies.
Comparison of three detection modes (interactance, reflectance and transmittance) was performed in terms of the necessity of visible (VIS) region and the interference caused by peel in SSC assessment of navel oranges. Spectra of 88 oranges with peel and peeled were collected using a commercial available CCD spectrometer. Partial least squares (PLS) regression was used to develop calibration models. Results showed that the participation of VIS region degraded the performances of PLS models in transmittance mode, while for the other two modes VIS–SWNIR models turned out to be the best. The peel of oranges did not bring about significant negative influence on SSC determination in any detection mode. The best calibration models were achieved with transmittance mode regardless of sample status. Future research should be focused on specific reasons such as the composition of orange peel and the influence of different sizes of fruits.
The feasibility of reflectance Vis/NIR spectroscopy was investigated for taste characterization of Valencia oranges based on taste attributes including soluble solids content (SSC) and titratable acidity (TA), as well as taste indices including SSC to TA ratio (SSC/TA) and BrimA. The robustness of multivariate analysis in terms of prediction was also assessed. Several combinations of various preprocessing techniques with moving average and Savitzky–Golay smoothing filters, standard normal variate (SNV) and multiplicative scatter correction (MSC) were used before calibration and the models were developed based on both partial least squares (PLS) and principle component regression (PCR) methods. The best models obtained with PLS method had root mean square errors of prediction (RMSEP) of 0.33 °Brix, 0.07%, 1.03 and 0.37, and prediction correlation coefficients (rp) of 0.96, 0.86, 0.87 and 0.92 for SSC, TA, SSC/TA, and BrimA, respectively. It was concluded that Vis/NIR spectroscopy combined with chemometrics could be an accurate and fast method for nondestructive prediction of taste attributes and indices of Valencia oranges. Moreover, the application of this technique was suggested for taste characterization, directly based on BrimA which is the best index related to fruit flavor rather than determination of SSC or TA alone.
This study proposes a new approach to discriminate low and full-fat yogurts using instrumental analysis and chemometric techniques. One hundred twenty six strawberry flavored yogurts were subjected to instrumental analysis of pH, color and firmness. Exploratory methods, such as Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA), and supervised classification methods, such as K-nearest neighbors (KNN), soft independent modeling of class analogy (SIMCA), and Partial Least Square Discriminant Analysis (PLSDA) were used for assessing the data. The results showed that low- and full-fat yogurts presented different with regard to all the variables analyzed. It was not possible to obtain total separation between the samples using PCA and HCA. KMN and PLSDA presented excellent performance toward the full-fat category, with 100% correct prediction which suggests only low-fat yogurts to be subjected to the traditional fat content determination methods. This approach can be incentivized by the health agencies aimed to optimize materials and financial resources.
The soluble solids content (SSC) and total acidity (TA) are the major characteristics for assessing quality and maturity of Nanfeng mandarin fruits. The feasibility of charge coupled device near infrared spectroscopy (CCD-NIRS) combining with effective wavelengths selection algorithm used to measure SSC and TA nondestructively was investigated. The effective wavelengths to SSC and TA were chosen by interval partial least squares (iPLS) in the wavelength range of 600–980nm. The predictive ability of SSC model used PLS regression was improved with r of 0.92 and RMSEP of 0.65°Brix using effective wavelengths of 681.36–740.51nm, 798.60–836.19nm and 945.52–962.75nm. The TA model was simplified with r of 0.64 and RMSEP of 0.09% using effective wavelengths of 817.57–836.19nm, 909.85–927.60nm and 945.52–962.75nm. The experimental results demonstrated that the CCD-NIRS technique combining with iPLS algorithm was a feasible method to measure SSC and TA of Nanfeng mandarin fruits nondestructively.
The aims of this study were (1) to map sensory attributes of vanilla ice cream with reduced fat and sugar, and (2) to determine drivers of liking by applying external preference mapping and reveal the relationship between descriptive attributes and hedonic judgments using the partial least squares method. Descriptive sensory profiles (n=11) and consumer test (n=117) of 6 samples of vanilla ice cream (3 traditional and 3 with reduced fat and sugar) were determined. The attributes brightness and sweet aftertaste for sample and creaminess (appearance and texture) and sweet aroma contributed positively to the acceptance of ice cream samples. The attributes aeration, powdered milk aroma and flavor, and white chocolate aroma and flavor contributed positively to the acceptance of the ice creams. The attributes hydrogenated fat aroma and flavor were responsible for the lower acceptance of samples. The reduction in fat and sugar did not necessarily cause a decrease in acceptance. The most important factors were selection of the appropriate sweetener system and the use of good quality raw material.
Two commercial portable spectrometers were compared for orange quality non-destructive predictions by developing partial least squares calibration models, reflectance mode spectra acquisition being used in both. One of them was a Vis/NIR spectrometer in which the radiation reflected by the fruit is collected and conducted by optic fiber to the three detectors (350–2500 nm) of the instrument. The other is an AOTF-NIR with a reflectance post-dispersive optical configuration and InGaAs (1100–2300 nm) detector. Four orange varieties were included in calibrations. The parameters studied were soluble solids content, acidity, titratable acidity, maturity index, flesh firmness, juice volume, fruit weight, rind weight, juice volume to fruit weight ratio, fruit colour index and juice colour index. The results indicate good performance of the predictive models, particularly for the direct NIR prediction of soluble solids content, and maturity index, the prediction of this last parameter being notable for its relevance and novelty. The RPD ratios for these parameters were in the range from 1.67 to 2.21 with the Labspec spectrometer, which showed better predictive performance, and from 1.03 to 2.33 with the Luminar instrument.
BACKGROUND: The vitamin C and polyphenol content of apples constitute quality and nutritional parameters of great interest for consumers, especially in terms of health. They are commonly measured using laborious reference methods. The purpose of this study was to evaluate the potential of near-infrared (NIR) spectroscopy as a rapid and non-destructive method to determine the sugar, vitamin C and total polyphenol content in apples.
RESULTS: The quality parameters of more than 150 apple genotypes were analyzed using NIR and reference methods. The results obtained using the least squares support vector machine regression method showed good to very good prediction performance. Low standard error of prediction values, in addition to relatively high ratio to prediction (RPD) values, demonstrated the precision of the models, especially for polyphenol and sugar content. High RPD values (5.1 and 4.3 for polyphenol and sugar, respectively) indicated that an accurate classification of apples based on their content could be achieved.
CONCLUSION: NIR spectroscopy coupled with the multivariate calibration technique could be used to accurately measure the quality parameters of apples. In addition, in the case of breeding programs, NIR spectroscopy can be considered an interesting tool for sorting varieties according to a range of concentrations. Copyright
Visible/near-infrared spectroscopy (Vis/NIRS) appears as a prominent technique for non-destructive fruit quality assessment. This research work was focused in to evaluate the use of Vis/NIRS in measuring the quality characteristics of intact Satsuma mandarin “Zaojin Jiaogan” (C. reticulata), and to establish the relationship between non-destructive Vis/NIR spectral measurements and the major physiological properties of fruit (firmness, soluble solids content (SSC) and acidity (pH)). Before calibration, two types of data pre-processing were used and NIR models were developed based on partial least square (PLS) and principal component regression (PCR) techniques. The prediction models indicated that a reasonable to excellent prediction performance could be expected for each property. The best SSC model had a mean square error of prediction (RMSEP) of 0.33 °Brix and correlation coefficient between predicted and measured values (r) of 0.94, the proposed model for the pH and the compression force had a RMSEP of 0.18 and 8.53, as well as, r of 0.8 and 0.83, respectively. It was concluded that by using the Vis/NIRS measurement technique, in the full spectral range (400–2350 nm), it is possible to assess the quality characteristics of mandarin.
The effect of cultivar, season, shelf-life and origin on the accuracy of near infrared (NIR) calibration models for the soluble solids content (SSC) and firmness of apple was studied based on a large spectral data set based on approximately 6000 apple fruit from different cultivars, origins, shelf-life exposure time and seasons. To interpret the variance in the spectra with respect to biological variability, functional analysis of variance (FANOVA) was used. From the FANOVA analysis it was concluded that the effects of cultivar, origin and shelf-life exposure time on the NIR spectra were all significant. The largest differences in the spectra were found around the water absorption peaks (970, 1170 and 1450 nm). External validations using independent data sets showed that the accuracy of the models increased considerably when more variability was included in the calibration data set. In general the RMSEP for predictions of the SSC were in the range 0.6–0.8 °Brix, while for Magness Taylor firmness it was 5.9–8.8 N, depending on the cultivar. It was shown that atypical data can lead to large validation errors. It is, therefore, important to collect a calibration data set which is sufficiently representative for future samples to be analyzed with the developed calibration models and to develop simple procedures for model adaptation during practical use.
Protein is an important component of milk powder. The fast and non-destructive detection of protein content in milk powder is important. Infrared spectroscopy technique was applied to achieve this purpose. Least-squares support vector machine (LS-SVM) was applied to building the protein prediction model based on spectral transmission rate. The determination coefficient for prediction ( was 0.981 and root mean square error for prediction (RMSEP) was 0.4115. It is concluded that infrared spectroscopy technique can quantify protein content in milk powder fast and non-destructively. The process is simple and easy to operate, and the prediction ability of LS-SVM is better than that of partial least square. Moreover, the comparison of prediction results showed that the performance of model with mid-infrared spectra data was better than that with near infrared spectra data.
A kernel PLS algorithm was implemented to estimate the sugar content of Golden Delicious apples based on NIR reflectance spectra in the range of 800–1690 nm. Covariance, Gaussian and polynomial kernel functions were considered. All kernels, except the covariance kernel, incorporate tuning parameters which were optimised by computer experiments. The calibration results were insensitive to the actual value of the tuning parameters over a wide range. No significant difference between the RMSEP values obtained with different kernels was obtained, irrespective of the applied transformation (none, log(1/R), Kubelka–Munck) or first order derivative calculation. A wavelet compression procedure was implemented to speed up the computation of the kernel Gram matrices. It was shown that the kernel Gram matrix computed with the approximation and detail coefficients of the wavelet transformed spectra converges in norm to the real kernel Gram matrix. The time required for calculating the kernel Gram matrix is inversely proportional to the compression ratio. It was shown that a compression ratio of up to 25 did not affect the accuracy of the kernel PLS calibration models.
Linear and nonlinear multivariate regressions were implemented to estimate sugar content of intact Gannan navel orange based on Vis–NIR diffuse reflectance spectroscopy in the wavelength range of 450–1750 nm. Four pre-processing methods, including average smoothing, multiplicative scatter correction (MSC), first and second derivatives, were applied to improve the predictive ability of the models. The models were developed by MLR, PCR, PLS, Poly-PLS and Spline-PLS with MSC pretreatment. Except MLR, the predictive results were insignificant among PCR, PLS, Poly-PLS and Spline-PLS by analysis of variance test at 5% level. The Spline-PLS model was superior to others with R of 0.87, RMSEP of 0.47∘Brix and SDR=2.34. The results illustrated Spline-PLS could be applied to deal with nonlinear problem, and Vis–NIR spectroscopy in combination with it, could determine sugar content of intact Gannan navel orange precisely.
Partial Least Squares (PLS) methods are particularly suited to the analysis of relationships between measures of brain activity and of behavior or experimental design. In neuroimaging, PLS refers to two related methods: (1) symmetric PLS or Partial Least Squares Correlation (PLSC), and (2) asymmetric PLS or Partial Least Squares Regression (PLSR). The most popular (by far) version of PLS for neuroimaging is PLSC. It exists in several varieties based on the type of data that are related to brain activity: behavior PLSC analyzes the relationship between brain activity and behavioral data, task PLSC analyzes how brain activity relates to pre-defined categories or experimental design, seed PLSC analyzes the pattern of connectivity between brain regions, and multi-block or multi-table PLSC integrates one or more of these varieties in a common analysis. PLSR, in contrast to PLSC, is a predictive technique which, typically, predicts behavior (or design) from brain activity. For both PLS methods, statistical inferences are implemented using cross-validation techniques to identify significant patterns of voxel activation. This paper presents both PLS methods and illustrates them with small numerical examples and typical applications in neuroimaging.
Nondestructive method of measuring soluble solids content (SSC) of citrus fruits was developed using Fourier transform near infrared reflectance (FT-NIR) measurements collected through optics fiber. The models describing the relationship between SSC and the NIR spectra of citrus fruits were developed and evaluated. Different spectra correction algorithms (standard normal variate (SNV), multiplicative signal correction (MSC)) were used in this study. The relationship between laboratory SSC and FT-NIR spectra of citrus fruits was analyzed via principle component regression (PCR) and partial least squares (PLS) regression method. Models based on the different spectral ranges were compared in this research. The first derivative and second derivative were applied to all spectra to reduce the effects of sample size, light scattering, instrument noise, etc. Different baseline correction methods were applied to improve the spectral data quality. Among them the second derivative method after baseline correction produced best noise removing capability and yielded optimal calibration models. A total of 170 NIR spectra were acquired; 135 NIR spectra were used to develop the calibration model; the remaining spectra were used to validate the model. The developed PLS model describing the relationship between SSC and NIR reflectance spectra could predict SSC of 35 samples with correlation coefficient of 0.995 and RMSEP of 0.79 degrees Brix.
The potential of visible and near-infrared reflectance spectroscopy (vis-NIRS) was investigated for its ability to nondestructively detect soluble solids contents (SSC) and pH in orange juices. A total of 104 orange juice samples were used for vis-NIRS at 325-1075 nm using a field spectroradiometer. Wavelet packet transform, standard normal variate transformation (SNV), and Savitzky-Golay first-derivative transformation were applied for the preprocessing of spectral data. The chemometrics of partial least-squares (PLS) regression analysis was performed on the processed spectral data. The evaluation of SSC and pH in orange juices by PLS regression with SNV showed the highest accuracy of the three preprocessing methods. The correlation coefficient (r), standard error of prediction, and the root-mean-square error of prediction for SSC were 0.98, 0.68, and 0.73, respectively, whereas those values for pH were 0.96, 0.06, and 0.06, respectively. The "fingerprint" representing features of orange juices or reflecting sensitivity to some elements at a certain band was proposed on the basis of regression coefficients. It is very useful in the field of food chemistry and further research on other materials. It is concluded that the vis-NIRS technique combined with chemometrics is promising for the fast and nondestructive detection of chemical components in orange juices or other materials.
maturity index (C) and vitamin C (D) against conventional measured parameters by LS-SVM regression using equatorial surface NIR spectra Non-destructive estimation of mandarin maturity status through portable VIS– NIR spectrophotometer
- F Pallottino
- F Paglia
- G Palma
- A Aquino
- S Menesatti
Fig. 4. Scatter plots of predicted TSS (A), TA (B), maturity index (C) and vitamin C (D) against conventional measured parameters by LS-SVM regression using equatorial
surface NIR spectra.
References
Antonucci, F., Pallottino, F., Paglia, G., Palma, A., D'Aquino, S., Menesatti, P., 2011.
Non-destructive estimation of mandarin maturity status through portable VIS–
NIR spectrophotometer. Food Bioprocess Technol. 4 (5), 809–813.