Comparison of machine learning models for classifying edible oils using Fourier‐transform infrared spectroscopy DOI
Hyeona Lim,

Seon Yeong Lee,

Jin Young Kim

и другие.

Bulletin of the Korean Chemical Society, Год журнала: 2024, Номер unknown

Опубликована: Дек. 23, 2024

Abstract Accurate classification and authentication of edible oils are essential for maintaining product quality, ensuring consumer safety, preserving market integrity. Therefore, this study aims to propose Fourier‐transform infrared (FT‐IR) spectroscopy, combined with advanced machine learning models, as a rapid non‐destructive technique classifying oils. The FT‐IR spectra seven oil types were analyzed across three spectral regions: the full range, C‐H stretching fingerprint region. Both absorbance second derivative used evaluate influence preprocessing on accuracy. Six models—principal component analysis followed by linear discriminant (PCA‐LDA), k‐nearest neighbors, decision tree, random forest, eXtreme Gradient Boosting, support vector machines (SVM)—were employed classify oils, achieving training accuracies 96.4%–100% testing 88.1%–100%. enhanced model performance improving resolution overlapping peaks, particularly in CH CO regions. Additionally, SHapley Additive exPlanations further revealed most critical features influencing predictions, offering valuable insights into decision‐making processes. This demonstrates effectiveness combining preprocessing, techniques findings highlight benefits enhancing superior PCA‐LDA SVM models. These results offer robust framework advancing emphasize potential artificial intelligence food quality control.

Язык: Английский

Machine Learning as a “Catalyst” for Advancements in Carbon Nanotube Research DOI Creative Commons
Guohai Chen, Dai‐Ming Tang

Nanomaterials, Год журнала: 2024, Номер 14(21), С. 1688 - 1688

Опубликована: Окт. 22, 2024

The synthesis, characterization, and application of carbon nanotubes (CNTs) have long posed significant challenges due to the inherent multiple complexity nature involved in their production, processing, analysis. Recent advancements machine learning (ML) provided researchers with novel powerful tools address these challenges. This review explores role ML field CNT research, focusing on how has enhanced research by (1) revolutionizing synthesis through optimization complex multivariable systems, enabling autonomous reducing reliance conventional trial-and-error approaches; (2) improving accuracy efficiency characterizations; (3) accelerating development applications across several fields such as electronics, composites, biomedical fields. concludes offering perspectives future potential integrating further into highlighting its driving forward.

Язык: Английский

Процитировано

5

Recent advances on artificial intelligence-based approaches for food adulteration and fraud detection in the food industry: Challenges and opportunities DOI
Puja Das, Ammar B. Altemimi, Pinku Chandra Nath

и другие.

Food Chemistry, Год журнала: 2024, Номер 468, С. 142439 - 142439

Опубликована: Дек. 10, 2024

Язык: Английский

Процитировано

5

Fuzzy Logic-based Barcode Scanning System for Food Products Halal Identification DOI
Nidhi Rajesh Mavani,

Mohamad Azri Ismail,

Norliza Abd Rahman

и другие.

Food Control, Год журнала: 2024, Номер unknown, С. 110926 - 110926

Опубликована: Сен. 1, 2024

Язык: Английский

Процитировано

4

Nondestructive Discrimination of Plant-Based Patty Containing Traditional Medicinal Roots Using Visible–Near-Infrared Hyperspectral Imaging and Machine Learning Techniques DOI Creative Commons
Ge Song,

Hwanjo Chung,

Reza Adhitama Putra Hernanda

и другие.

Chemosensors, Год журнала: 2025, Номер 13(5), С. 158 - 158

Опубликована: Апрель 25, 2025

The interest in traditional meat being replaced by plant-based food has increased throughout the years. Some agricultural products, such as root crops, could be incorporated into alternative products due to health benefits. However, relevant studies have discovered that some roots are considered allergen materials, necessitating further identification maintain consumer safety. Aside from high accuracy, limitations offered methods a reason employ nondestructive methods. This study aimed develop hyperspectral imaging system measuring 400 nm 1000 spectral range for of soybean-based patty. Four thin-sliced medicinal (tianma (Gastrodia elata), balloon flower (Platycodon grandiflorum), deodeok (Codonopsis lanceolata), and ginseng (Panax ginseng)) were patty with concentration 5% w/w. Moreover, support vector machine (SVM) learning one-dimensional convolutional neural networks (1D-CNN) realized discrimination model tandem data extracted image. Our demonstrated SVM effectively discriminates between original addition, an F1-score, precision, recall beyond 96.77%. optimum was achieved using standard normal variate (SNV) spectra.

Язык: Английский

Процитировано

0

Optimization of DNA Extraction from Fish Oil Residuals Based on Magnetic Bead Method DOI
Wei Zhao,

Qinting Jiang,

Xiaoling Zhou

и другие.

Research Square (Research Square), Год журнала: 2025, Номер unknown

Опубликована: Май 2, 2025

Abstract As a high-value functional food, fish oil has substantial market demand, yet adulteration issues are frequent. PCR-based detection of species-specific genes is currently the most direct and reliable molecular approach for identifying source species their relative abundance in analysis. The concentration purity DNA critical ensuring efficient amplification accurate detection. Therefore, this study optimized extraction protocol from using magnetic bead-based methods by refining key parameters, including bead type, binding buffer composition, particle size, volume, concentration, washing conditions, elution conditions. optimal conditions were determined as follows: silica-coated OH-500 beads, containing 3 mol/L guanidine isothiocyanate, 50 µL 70% ethanol washing, water at 56°C 15 minutes. Compared to commercial kits, method improved efficiency nearly 10% while demonstrating high reproducibility (CV < 5%). This refined provides an technical solution oil.

Язык: Английский

Процитировано

0

Comparison of Fuel Properties of Alternative Fuels from Insect Lipids and Their Blending with Diesel Fuel DOI Open Access
Ji Eun Lee, Hyun Sung Jang,

Yeo Jin Yun

и другие.

Sustainability, Год журнала: 2025, Номер 17(10), С. 4295 - 4295

Опубликована: Май 9, 2025

Drop-in fuels are renewable alternatives that can be integrated into an existing fuel infrastructure without modification. Among these, synthesized from hydroprocessed lipids have garnered significant attention owing to their compatibility with petroleum-based diesel. In this study, we investigated the feasibility of hydrodeoxygenated insect oil (HIO), derived black soldier fly larvae (Hermetia illucens; BSFL), as a drop-in for diesel blend. The optimal growth conditions BSFL were studied maximize lipid production, and extracted was subjected hydrodeoxygenation (HDO) via catalytic reaction. HIO blended commercial at ratios 5–30%, its properties compared A detailed property analysis conducted 5% blend evaluate suitability fuel. Characterization fuels’ physicochemical carried out assess potential insect-derived applications.

Язык: Английский

Процитировано

0

Jackfruit-like ZnO gas sensor for monitoring ethyl formate emissions from fish meal DOI
Zhaopeng Li, Pei Li,

Peisi Yin

и другие.

Sensor Review, Год журнала: 2025, Номер unknown

Опубликована: Апрель 7, 2025

Purpose This paper aims to develop a highly sensitive resistive gas sensor for accurately detecting ethyl formate achieve reliable and real-time monitoring of fish meal spoilage. Design/methodology/approach Based on the one-step solvothermal reduction method, in specific triethylene glycol solution environment high temperature, 3D ZnO sensing material with jackfruit-like structure was prepared meal. Findings The based displays response (69.68–100 ppm) at 280°C 43% RH good (12.18–100 80% RH, ultra-low detection limit 10 ppb excellent selectivity, repeatability long-term stability. mechanism is due gain or loss electrons caused by surface reaction. unique structure, abundant oxygen vacancies large area may be another factor contributing its performance. Originality/value authors first developed an sensor, results were compared previously published data. analysis showed demonstrated work highlights potential sensors evaluate quality.

Язык: Английский

Процитировано

0

Rapid and noncontact identification of soybean flour in edible insect using NIR spectral imager: A case study in Protaetia brevitarsis seulensis powder DOI
Reza Adhitama Putra Hernanda, Juntae Kim, Mohammad Akbar Faqeerzada

и другие.

Food Control, Год журнала: 2024, Номер 169, С. 111019 - 111019

Опубликована: Ноя. 8, 2024

Язык: Английский

Процитировано

1

Comparison of machine learning models for classifying edible oils using Fourier‐transform infrared spectroscopy DOI
Hyeona Lim,

Seon Yeong Lee,

Jin Young Kim

и другие.

Bulletin of the Korean Chemical Society, Год журнала: 2024, Номер unknown

Опубликована: Дек. 23, 2024

Abstract Accurate classification and authentication of edible oils are essential for maintaining product quality, ensuring consumer safety, preserving market integrity. Therefore, this study aims to propose Fourier‐transform infrared (FT‐IR) spectroscopy, combined with advanced machine learning models, as a rapid non‐destructive technique classifying oils. The FT‐IR spectra seven oil types were analyzed across three spectral regions: the full range, C‐H stretching fingerprint region. Both absorbance second derivative used evaluate influence preprocessing on accuracy. Six models—principal component analysis followed by linear discriminant (PCA‐LDA), k‐nearest neighbors, decision tree, random forest, eXtreme Gradient Boosting, support vector machines (SVM)—were employed classify oils, achieving training accuracies 96.4%–100% testing 88.1%–100%. enhanced model performance improving resolution overlapping peaks, particularly in CH CO regions. Additionally, SHapley Additive exPlanations further revealed most critical features influencing predictions, offering valuable insights into decision‐making processes. This demonstrates effectiveness combining preprocessing, techniques findings highlight benefits enhancing superior PCA‐LDA SVM models. These results offer robust framework advancing emphasize potential artificial intelligence food quality control.

Язык: Английский

Процитировано

1