Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine
Traditional herbal medicine is an important part of the global health system with persistent issues like adulteration and misidentification. While current global standards employ chromatographic identification to combat this issue, there are some disadvantages with such methods. This study looked at...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Language: | English |
Published: |
2022
|
Subjects: | |
Online Access: | http://eprints.usm.my/60131/1/24%20Pages%20from%20YEAP%20ZHAO%20QIN.pdf http://eprints.usm.my/60131/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my.usm.eprints.60131 |
---|---|
record_format |
eprints |
spelling |
my.usm.eprints.60131 http://eprints.usm.my/60131/ Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine Yeap, Zhao Qin RS1-441 Pharmacy and materia medica Traditional herbal medicine is an important part of the global health system with persistent issues like adulteration and misidentification. While current global standards employ chromatographic identification to combat this issue, there are some disadvantages with such methods. This study looked at five traditional herbal medicine, namely Anoectochilus roxburghii, Aristolochia manshuriensis, Dioscorea hamiltonii, Gelsemium elegans and Alisma orientalis, and ways to classify them. Infrared spectra of the herbs were collected from a total of 200 samples from 20 °C to 120 °C at 10 °C intervals. Infrared signals of functional groups on the main chemical constituents for every herb were present in the infrared spectra collected. Principal component analysis of these spectra found a limitation where the success of the analysis might require less types of herbs included. Computational combination of thermally perturbed spectra was performed to obtain two-dimensional chemical fingerprints for every sample. Machine Learning Classifier were trained to generate models. The model classified the herbs with an accuracy of 87.9 % when all the herbs were included in training. When Alisma orientalis was excluded from training to measure the robustness of the model, 91.3 % of the samples were classified correctly. 2022-11 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/60131/1/24%20Pages%20from%20YEAP%20ZHAO%20QIN.pdf Yeap, Zhao Qin (2022) Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine. Masters thesis, Perpustakaan Hamzah Sendut. |
institution |
Universiti Sains Malaysia |
building |
Hamzah Sendut Library |
collection |
Institutional Repository |
continent |
Asia |
country |
Malaysia |
content_provider |
Universiti Sains Malaysia |
content_source |
USM Institutional Repository |
url_provider |
http://eprints.usm.my/ |
language |
English |
topic |
RS1-441 Pharmacy and materia medica |
spellingShingle |
RS1-441 Pharmacy and materia medica Yeap, Zhao Qin Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
description |
Traditional herbal medicine is an important part of the global health system with persistent issues like adulteration and misidentification. While current global standards employ chromatographic identification to combat this issue, there are some disadvantages with such methods. This study looked at five traditional herbal medicine, namely Anoectochilus roxburghii, Aristolochia manshuriensis, Dioscorea hamiltonii, Gelsemium elegans and Alisma orientalis, and ways to classify them. Infrared spectra of the herbs were collected from a total of 200 samples from 20 °C to 120 °C at 10 °C intervals. Infrared signals of functional groups on the main chemical constituents for every herb were present in the infrared spectra collected. Principal component analysis of these spectra found a limitation where the success of the analysis might require less types of herbs included. Computational combination of thermally perturbed spectra was performed to obtain two-dimensional chemical fingerprints for every sample. Machine Learning Classifier were trained to generate models. The model classified the herbs with an accuracy of 87.9 % when all the herbs were included in training. When Alisma orientalis was excluded from training to measure the robustness of the model, 91.3 % of the samples were classified correctly. |
format |
Thesis |
author |
Yeap, Zhao Qin |
author_facet |
Yeap, Zhao Qin |
author_sort |
Yeap, Zhao Qin |
title |
Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
title_short |
Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
title_full |
Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
title_fullStr |
Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
title_full_unstemmed |
Using Mid-Infrared Spectroscopic Fingerprinting, Multivariate Analysis And Machine Learning To Differentiate Traditional Herbal Medicine |
title_sort |
using mid-infrared spectroscopic fingerprinting, multivariate analysis and machine learning to differentiate traditional herbal medicine |
publishDate |
2022 |
url |
http://eprints.usm.my/60131/1/24%20Pages%20from%20YEAP%20ZHAO%20QIN.pdf http://eprints.usm.my/60131/ |
_version_ |
1794552247383752704 |
score |
13.222552 |