Hyperspectral Imaging and Support Vector Machine: A Powerful Combination to Differentiate Black Cohosh (Actaea racemosa) from Other Cohosh Species

Abstract

(black cohosh) has a history of traditional use in the treatment of general gynecological problems. However, the plant is known to be vulnerable to adulteration with other cohosh species. This study evaluated the use of shortwave infrared hyperspectral imaging (SWIR-HSI) in tandem with chemometric data analysis as a fast alternative method for the discrimination of four cohosh species () and 36 commercial products labelled as black cohosh. The raw material and commercial products were analyzed using SWIR-HSI and ultra-high-performance liquid chromatography coupled to mass spectrometry (UHPLC-MS) followed by chemometric modeling. From SWIR-HSI data (920 - 2514 nm), the range containing the discriminating information of the four species was identified as 1204 - 1480 nm using Matlab software. After reduction of the data set range, partial least squares discriminant analysis (PLS-DA) and support vector machine discriminant analysis (SVM-DA) models with coefficients of determination ( ) of ≥ 0.8 were created. The novel SVM-DA model showed better predictions and was used to predict the commercial product content. Seven out of 36 commercial products were recognized by the SVM-DA model as being true black cohosh while 29 products indicated adulteration. Analysis of the UHPLC-MS data demonstrated that six commercial products could be authentic black cohosh. This was confirmed using the fragmentation patterns of three black cohosh markers (cimiracemoside C; 12-,21-dihydroxycimigenol-3--L-arabinoside; and 24--acetylhydroshengmanol-3---D-xylopyranoside). SWIR-HSI in conjunction with chemometric tools (SVM-DA) could identify 80% adulteration of commercial products labelled as black cohosh.

Spectrometric Classification of Bamboo Shoot Species by Comparison of Different Machine Learning Methods

“…In this study, a high range of OVA.ACC from 0.91 to 0.99 was found using the Det+2nd spectra and SVM model. The SVM model performed better than PLSDA and RF, similar to the finding of Tankeu et al (2018), who reported that by using of hyperspectral imaging data, the SVM model yielded better predictions for the identification of true black cohosh, and Sun et al (2017), who showed that the SVM model combined with a hyperspectral reflectance imaging technique had high accuracy (92.96-97.28%) in the detection of chilling peaches. SVM combined with NIR spectroscopy also yielded a promising result for the prediction of the chilling storage stage of eggplants (Tsouvaltzis et al 2020) and different rice flour types (Sampaio et al 2020).…”

A Hyperspectral Imaging Approach for Classifying Geographical Origins of Rhizoma Atractylodis Macrocephalae Using the Fusion of Spectrum-Image in VNIR and SWIR Ranges (VNIR-SWIR-FuSI)

“…In this work, supervised pattern recognition models were adopted for the origin identification of RAM slices. There have been a variety of models available for classification, including partial least square-discriminant analysis (PLS-DA) [34,35], linear discriminate analysis (LDA) [36], support vector machine (SVM) [8] and back propagation neural network (BPNN) [5]. PLS-DA and SVM were selected herein.…”

“…However, these studies focus on HSI systems within the visible and near-infrared (VNIR, about 400–1000 nm) range, or the short-wave near-infrared (SWIR, about 900–1700 nm) range. Moreover, there have been few HSI studies on comparing the classification performance of spectrum and image fusion with VNIR and SWIR range combined, especially in the realm of traditional Chinese medicine (TCM) due to lack of appropriate methodology [8,9,10].…”

Botanical ingredient identification and quality assessment: strengths and limitations of analytical techniques

“…Even though the assessment of plant material is prone to errors with regard to moisture, degree of comminution, analysis of extracts or crude material etc., the number of applications in plant quality control increases. There are reports on the differentiation between black cohosh species (Tankeu et al 2018), determination of geographic origin of Chinese Astragalus samples (Zhang and Nie 2010), quality assessment of South African herbal tea blends (Djokam et al 2017;Sandasi et al 2018), quantification of constituents (Mavimbela et al 2018;Pezzei et al 2018), and most recently even prediction on activities of plant extracts (Nikzad-Langerodi et al 2017;Pezzei et al 2018;Schönbichler et al 2014). However, reliability of IR-based results strongly depends on well-developed reference libraries that integrate closely related and potentially adulterating species to ensure that the spectra obtained is accurate to species.…”