Determination of protein and fatty acid composition of shell-intact upland cottonseed using near-infrared reflectance spectroscopy

Near-Infrared Spectroscopy Analysis of the Phytic Acid Content in Fuzzy Cottonseed Based on Machine Learning Algorithms

“…These problems can be solved by using SNV and 1D preprocessing. This may be one of the reasons why the most suitable preprocessing method after screening is the combination of SNV+1D [53]. Notwithstanding, there are four techniques that reduced the predictive performance of the model, and there are possible reasons behind it.…”

Machine learning applications for multi-source data of edible crops: A review of current trends and future prospects

“…Relatively speaking, the clustering effect of taste sensor data was better. The clustering trend of PCA models based on data fusion was also unsatisfactory ( Chen et al, 2015 ) B Palm oil HPLC-UV, HPLC-CAD Both can provide sample information in a non-selective manner, and the fingerprint can serve as a complete analytical data Using PCA to visualize samples of HPLC-CAD and HPLC-UV, two outliers were found in HPLC-CAD, while there were no outliers in HPLC-UV ( Obisesan et al, 2017 ) D Saffron NIR, MIR Both are easy to operate, fast, and environmentally friendly, but they are selective, so in order to overcome their shortcomings, chemometrics is needed NIR-PCA showed two trends in the distribution of saffron samples, and MIR-PCA showed no significant distribution trend compared to NIR ( Amirvaresi et al, 2021 ) PLSR D Olive oil NIR, MIR – NIR: R 2 = 0.896, RMSEP = 7.09; MIR: R 2 = 0.966, RMSEP = 4.04; LLF: R 2 = 0.975, RMSEP = 3.44; HLF: R 2 = 0.988, RMSEP = 2.86 (Best) ( Li et al, 2019 ) E Ziziphus jujuba NIR, MIR – NIR: R 2 = 0.9312, RPD = 2.82 MIR: R 2 = 0.8951, RPD = 2.28 LLF: R 2 = 0.9475, RPD = 2.10 MLF: R 2 = 0.9621, RPD = 2.44 (Best) ( Arslan et al, 2019 ) E Cottonseed NIR, GC–MS NIR is high-throughput, simple and low-cost R 2 cal > 0.7 ( Zhuang et al, 2023 ) SVR E Yuezhou Longjing tea NIR, HPLC NIR has the advantages of non-destructive testing, fast testing speed, and high efficiency Sensory quality: RPD(PLSR) = 1.888,RPD (RF) = 2.033, RPD(SVR) = 2.485 (Best); Catechins: RPD(PLSR) = 1.857, RPD(SVR) = 2.088, RPD(RF) = 2.584 (Best); Caffeine: RPD(PLSR) = 2.076, RPD(SVR) = 2.799, RPD(RF) = 2.873 (Best) ( Jia et al, 2022 ) E Ginkgo biloba leaf extract NIR, HPLC ...…”

“…Cottonseed is one of the important oil-bearing crops, and its important quality evaluation indicators are fatty acid and protein. Zhuang et al (2023) obtained the phenotypic data of 17 fatty acids, oil, and proteins in shell-intact upland cottonseed using GC–MS and Soxhlet extraction methods, and correlated the content with the preprocessed NIR datasets through PLSR. The NIR spectral region of 950–1650 nm was selected as the model input data because fatty acids and proteins have strong absorption peaks in this range.…”