Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.

Accurate monitoring of glucose levels is essential in the field of diabetes detection and prevention to ensure appropriate treatment planning. Conventional blood glucose monitoring methods, although widely used, are intrusive and frequently result in discomfort. This study investigates the use of Ra...

Full description

Saved in:
Bibliographic Details
Main Authors: Tri Ngo Quang, Tung Nguyen Thanh, Duc Le Anh, Huong Pham Thi Viet, Doanh Sai Cong
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2024-01-01
Series:PLoS ONE
Online Access:https://doi.org/10.1371/journal.pone.0311482
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846129372252078080
author Tri Ngo Quang
Tung Nguyen Thanh
Duc Le Anh
Huong Pham Thi Viet
Doanh Sai Cong
author_facet Tri Ngo Quang
Tung Nguyen Thanh
Duc Le Anh
Huong Pham Thi Viet
Doanh Sai Cong
author_sort Tri Ngo Quang
collection DOAJ
description Accurate monitoring of glucose levels is essential in the field of diabetes detection and prevention to ensure appropriate treatment planning. Conventional blood glucose monitoring methods, although widely used, are intrusive and frequently result in discomfort. This study investigates the use of Raman spectroscopy as a non-invasive method for estimating glucose concentrations. Our proposition entails employing machine learning models to categorize glucose levels by utilizing Raman spectrum data. The collection consists of deionized water samples containing glucose with defined amounts, guaranteeing great purity and little interference. We assess the efficacy of three machine learning models in categorizing glucose levels which including Extra Trees, Random Forest, and Support Vector Machine (SVM). In addition, we employ data pre-processing techniques such as fluorescence background removal and hotspot series extraction to improve the performance of the model. The primary results demonstrate that the utilization of these pre-processing techniques greatly enhances the accuracy of classification. Among these techniques, the Extra Trees model achieves the highest accuracy, reaching 95%. This study showcases the viability of employing machine learning techniques to forecast glucose levels based on Raman spectroscopy data. Additionally, it emphasizes the significance of data pre-processing in enhancing the accuracy of the model's results.
format Article
id doaj-art-2c9e16021e8444c6b329f0e8535ee5ed
institution Kabale University
issn 1932-6203
language English
publishDate 2024-01-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS ONE
spelling doaj-art-2c9e16021e8444c6b329f0e8535ee5ed2024-12-10T05:32:28ZengPublic Library of Science (PLoS)PLoS ONE1932-62032024-01-011912e031148210.1371/journal.pone.0311482Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.Tri Ngo QuangTung Nguyen ThanhDuc Le AnhHuong Pham Thi VietDoanh Sai CongAccurate monitoring of glucose levels is essential in the field of diabetes detection and prevention to ensure appropriate treatment planning. Conventional blood glucose monitoring methods, although widely used, are intrusive and frequently result in discomfort. This study investigates the use of Raman spectroscopy as a non-invasive method for estimating glucose concentrations. Our proposition entails employing machine learning models to categorize glucose levels by utilizing Raman spectrum data. The collection consists of deionized water samples containing glucose with defined amounts, guaranteeing great purity and little interference. We assess the efficacy of three machine learning models in categorizing glucose levels which including Extra Trees, Random Forest, and Support Vector Machine (SVM). In addition, we employ data pre-processing techniques such as fluorescence background removal and hotspot series extraction to improve the performance of the model. The primary results demonstrate that the utilization of these pre-processing techniques greatly enhances the accuracy of classification. Among these techniques, the Extra Trees model achieves the highest accuracy, reaching 95%. This study showcases the viability of employing machine learning techniques to forecast glucose levels based on Raman spectroscopy data. Additionally, it emphasizes the significance of data pre-processing in enhancing the accuracy of the model's results.https://doi.org/10.1371/journal.pone.0311482
spellingShingle Tri Ngo Quang
Tung Nguyen Thanh
Duc Le Anh
Huong Pham Thi Viet
Doanh Sai Cong
Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
PLoS ONE
title Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
title_full Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
title_fullStr Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
title_full_unstemmed Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
title_short Classification of glucose-level in deionized water using machine learning models and data pre-processing technique.
title_sort classification of glucose level in deionized water using machine learning models and data pre processing technique
url https://doi.org/10.1371/journal.pone.0311482
work_keys_str_mv AT tringoquang classificationofglucoselevelindeionizedwaterusingmachinelearningmodelsanddatapreprocessingtechnique
AT tungnguyenthanh classificationofglucoselevelindeionizedwaterusingmachinelearningmodelsanddatapreprocessingtechnique
AT ducleanh classificationofglucoselevelindeionizedwaterusingmachinelearningmodelsanddatapreprocessingtechnique
AT huongphamthiviet classificationofglucoselevelindeionizedwaterusingmachinelearningmodelsanddatapreprocessingtechnique
AT doanhsaicong classificationofglucoselevelindeionizedwaterusingmachinelearningmodelsanddatapreprocessingtechnique