In Vitro Glucose Measurement from NIR and MIR Spectroscopy - Comprehensive Benchmark of Machine Learning and Filtering Chemometrics

Aug 4, 2026

Glucose quantitative analysis via spectroscopy remains crucial for both scientific research and industrial deployment. A key challenge is selecting effective preprocessing and regression tools before modeling. This study performs a comprehensive comparative evaluation of machine learning models and preprocessing filtering methods applied to near-infrared, mid-infrared, and combined NIR–MIR spectra. Spectral data from glucose solutions were processed using moving average, Savitzky–Golay, multiplicative scatter correction, and normalization, then modeled with linear, traditional nonlinear, and artificial neural network approaches. Results show linear models outperform nonlinear models, while neural networks achieve comparable accuracy. Overall best performance is obtained with convolutional moving average and Savitzky–Golay filters, supporting improved glucose quantification technologies.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Glucose monitoring and spectroscopy role

Machine learning approaches for NIR/MIR glucose quantification

Heliyon 10 (2024) e30981 Contents lists available at ScienceDirect Heliyon journal [homepage: www.cell.com/heliyon](homepage: www.cell.com/heliyon) | Research article
In Vitro Glucose Measurement from NIR and MIR Spectroscopy: Comprehensive Benchmark of Machine Learning and
Filtering Chemometrics | | | | --- | --- | --- | | Heydar Khadema, b, c, 1, *, Hoda Nemata, Jackie Elliott d, e, 1, Mohammed Benaissaa
a Department of Electronic and Electrical Engineering, University of Sheffield, UK b Department of Computer Science, University of Manchester, Manchester, UK c Artificial Intelligence & Machine Learning Team, KultraLab, London, UK d Department of Oncology and Metabolism, University of Sheffield, UK
e Sheffield Teaching Hospitals, Diabetes and Endocrine Centre, Northern General Hospital, Sheffield, UK | | | | A R T I C L E I N F O
Keywords:
Near-infrared
Mid-infrared Spectroscopy Machine learning Artificial intelligence Glucose
Signal processing | A B S T R A C T
The quantitative analysis of glucose using spectroscopy is a topic of great significance and interest in science and industry. One conundrum in this area is deploying appropriate preprocessing and regression tools. To contribute to addressing this challenge, in this study, we conducted a comprehensive and novel comparative analysis of various machine learning and preprocessing filtering techniques applied to near-infrared, mid-infrared, and a combination of near-infrared and mid-infrared spectroscopy for glucose assay. Our objective was to evaluate the effectiveness of these techniques in accurately predicting glucose levels and to determine which approach was most optimal. Our investigation involved the acquisition of spectral data from samples of glucose solutions using the three aforementioned spectroscopy techniques. The data was subjected to several preprocessing filtering methods, including convolutional moving average, Savitzky-Golay, multiplicative scatter correction, and normalisation. We then applied representative machine learning algorithms from three categories: linear modelling, traditional nonlinear modelling, and artificial neural networks. The evaluation results revealed that linear models exhibited higher predictive accuracy than nonlinear models, whereas artificial neural network models demonstrated comparable performance. Additionally, the comparative analysis of various filtering methods demonstrated that the convolutional moving average and Savitzky-Golay filters yielded the most precise outcomes overall. In conclusion, our study provides valuable insights into the efficacy of different machine learning techniques for glucose measurement and highlights the importance of applying appropriate filtering methods in enhancing predictive accuracy. These findings have important implications for the development of new and improved glucose quantification technologies. | | * Corresponding author. Department of Electronic and Electrical Engineering, University of Sheffield, UK. [E-mail address:](E-mail address: h.khadem@sheffield.ac.uk)[ h.khadem@sheffield.ac.uk](E-mail address: h.khadem@sheffield.ac.uk) (H. Khadem). 1 This article is based on research conducted at the University of Sheffield. [https://doi.org/10.1016/j.heliyon.2024.e30981](https://doi.org/10.1016/j.heliyon.2024.e30981) Received 5 May 2024; Received in revised form 8 May 2024; Accepted 8 May 2024 Available online 9 May 2024 2405-8440/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license ([http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/)). 1. Introduction Glucose monitoring is an example of a promising area of research with a wide variety of applications in different fields. With the help of machine learning, it has become promising to accurately predict glucose levels using methods such as near-infrared (NIR) and mid-infrared (MIR) spectroscopy. Th

FAQs

What preprocessing filtering methods were evaluated for glucose spectral data?

The study compares convolutional moving average, Savitzky–Golay, multiplicative scatter correction, and normalization before regression modeling.

Which model categories were used to predict glucose levels from NIR/MIR spectra?

Models span linear modeling, traditional nonlinear modeling, and artificial neural networks, including representative algorithms from each category.

What were the main findings about prediction accuracy across model types and filters?

Linear models achieved higher accuracy than nonlinear models, while artificial neural networks showed comparable performance. Convolutional moving average and Savitzky–Golay filters produced the most precise overall outcomes.

Aug 4, 2026

Upload by Genevieve