Pub Date : 2025-03-08DOI: 10.1016/j.chemolab.2025.105365
Sojeong Bae , Ku Kang , Young Kyun Kim , Yoon Jeong Jang , Doo-Hee Lee
Chemical warfare agents (CWAs) pose serious risks, requiring rapid, accurate detection. This study presents a real-time, lightweight AI system using YOLOv8 and colorimetric sensors, designed for field deployment. A dataset of 1,340 images captured under varying conditions enhances robustness. The model achieves 91.3% [email protected] and 10.4 ms/frame inference time on portable hardware. This system bridges the gap between laboratory methods and scalable field detection, ensuring efficient, on-site CWA identification for military, emergency response, and public health applications.
{"title":"Field-deployable real-time AI System for chemical warfare agent detection using YOLOv8 and colorimetric sensors","authors":"Sojeong Bae , Ku Kang , Young Kyun Kim , Yoon Jeong Jang , Doo-Hee Lee","doi":"10.1016/j.chemolab.2025.105365","DOIUrl":"10.1016/j.chemolab.2025.105365","url":null,"abstract":"<div><div>Chemical warfare agents (CWAs) pose serious risks, requiring rapid, accurate detection. This study presents a real-time, lightweight AI system using YOLOv8 and colorimetric sensors, designed for field deployment. A dataset of 1,340 images captured under varying conditions enhances robustness. The model achieves 91.3% [email protected] and 10.4 ms/frame inference time on portable hardware. This system bridges the gap between laboratory methods and scalable field detection, ensuring efficient, on-site CWA identification for military, emergency response, and public health applications.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"261 ","pages":"Article 105365"},"PeriodicalIF":3.7,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143592841","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-05DOI: 10.1016/j.chemolab.2025.105367
Mingwei Jia , Qiao Liu , Lingwei Jiang , Yi Liu , Zengliang Gao , Tao Chen
Deep learning-based just-in-time soft sensors effectively handle the strong nonlinearity of complex process industry, but their implementation faces significant challenges in interpretability and time cost. Hence, a just-in-time soft sensor based on spatiotemporal graph decoupling is proposed. To decrease time cost, it employs a global-local modeling strategy: pre-training on all historical data to build a global model, and fine-tuning with relevant samples to deliver a local model. To enhance interpretability, couplings that reflect how variables interact with each other in spatiotemporal dimensions are constructed, conforming to prior knowledge, to guide the graph neural network as a global model during pre-training. The global model decouples variables to quantify their influence as intrinsic information, enabling a clearer understanding of how each variable contributes to the prediction. Following the intrinsic information, relevant samples are then selected with the preset relevance metric to fine-tune the global model. Finally, two industrial cases demonstrate this model's low runtime, effectiveness, and physical consistency from the perspectives of underlying physics.
{"title":"Just-in-time process soft sensor with spatiotemporal graph decoupled learning","authors":"Mingwei Jia , Qiao Liu , Lingwei Jiang , Yi Liu , Zengliang Gao , Tao Chen","doi":"10.1016/j.chemolab.2025.105367","DOIUrl":"10.1016/j.chemolab.2025.105367","url":null,"abstract":"<div><div>Deep learning-based just-in-time soft sensors effectively handle the strong nonlinearity of complex process industry, but their implementation faces significant challenges in interpretability and time cost. Hence, a just-in-time soft sensor based on spatiotemporal graph decoupling is proposed. To decrease time cost, it employs a global-local modeling strategy: pre-training on all historical data to build a global model, and fine-tuning with relevant samples to deliver a local model. To enhance interpretability, couplings that reflect how variables interact with each other in spatiotemporal dimensions are constructed, conforming to prior knowledge, to guide the graph neural network as a global model during pre-training. The global model decouples variables to quantify their influence as intrinsic information, enabling a clearer understanding of how each variable contributes to the prediction. Following the intrinsic information, relevant samples are then selected with the preset relevance metric to fine-tune the global model. Finally, two industrial cases demonstrate this model's low runtime, effectiveness, and physical consistency from the perspectives of underlying physics.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"261 ","pages":"Article 105367"},"PeriodicalIF":3.7,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143579500","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-23DOI: 10.1016/j.chemolab.2025.105361
Zihang Wang , Shuai Li , Xiaofeng Zhou , Shijie Zhu
Time series causal discovery aims to identify cause-effect relationships among variables from time series data, providing valuable insights into complex real-world scenarios. However, existing constraint-based causal discovery methods face challenges such as limited detection power, stemming from issues like dimensionality explosion and uncertainty caused by indirect paths. To address these problems, we propose a novel iterative conditional variable selection method designed for lagged, linear, and nonlinear causal discovery in time series. (1) Firstly, we block indirect information while minimizing the dimensionality of the conditioning set. Specifically, our method selects the parent set of each target variable as the conditioning set, which includes only those variables involved in the indirect path. (2) Then, we refine the conditioning set by selecting a subset of the parent set for each target variable to focus on indirect causal relationships. (3) Finally, the iterative application of steps (1) and (2) progressively corrects the indirect paths, leading to a significant improvement in detection power. Experimental results on synthetic and public datasets, as well as for varying time lags, node counts, and a chemical fault diagnosis case, demonstrate that our method outperforms state-of-the-art (SOTA) approaches.
{"title":"An iterative conditional variable selection method for constraint-based time series causal discovery","authors":"Zihang Wang , Shuai Li , Xiaofeng Zhou , Shijie Zhu","doi":"10.1016/j.chemolab.2025.105361","DOIUrl":"10.1016/j.chemolab.2025.105361","url":null,"abstract":"<div><div>Time series causal discovery aims to identify cause-effect relationships among variables from time series data, providing valuable insights into complex real-world scenarios. However, existing constraint-based causal discovery methods face challenges such as limited detection power, stemming from issues like dimensionality explosion and uncertainty caused by indirect paths. To address these problems, we propose a novel iterative conditional variable selection method designed for lagged, linear, and nonlinear causal discovery in time series. (1) Firstly, we block indirect information while minimizing the dimensionality of the conditioning set. Specifically, our method selects the parent set of each target variable as the conditioning set, which includes only those variables involved in the indirect path. (2) Then, we refine the conditioning set by selecting a subset of the parent set for each target variable to focus on indirect causal relationships. (3) Finally, the iterative application of steps (1) and (2) progressively corrects the indirect paths, leading to a significant improvement in detection power. Experimental results on synthetic and public datasets, as well as for varying time lags, node counts, and a chemical fault diagnosis case, demonstrate that our method outperforms state-of-the-art (SOTA) approaches.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"260 ","pages":"Article 105361"},"PeriodicalIF":3.7,"publicationDate":"2025-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143520783","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-22DOI: 10.1016/j.chemolab.2025.105353
Yanan Zhang , Gaowei Yan , Shuyi Xiao , Fang Wang , Guanjia Zhao , Suxia Ma
In process industries, the complexity and variability of working conditions make it challenging to accurately measure product quality. While data-driven models have developed rapidly, they often overlook the underlying physical or chemical mechanisms. To address this, we propose a hybrid modeling approach that combines mechanism- and data-driven methods. Historical and current working condition data are processed through a hidden layer to extract features. The partial differential equation is discretized and approximated using the forward Euler method to derive mechanism-based quality variable values. These values are then combined with real data through a weighted mix to create a new label for dynamic regression. Additionally, a domain adaptation regularization term is introduced to align the distributions of different working conditions. Through analyses of three process industry datasets, we demonstrate that this method can predict unmeasurable variables with reasonable accuracy and exhibits stronger generalization ability compared to pure data-driven models.
{"title":"Mechanism- and data-driven based dynamic hybrid modeling for multi-condition processes","authors":"Yanan Zhang , Gaowei Yan , Shuyi Xiao , Fang Wang , Guanjia Zhao , Suxia Ma","doi":"10.1016/j.chemolab.2025.105353","DOIUrl":"10.1016/j.chemolab.2025.105353","url":null,"abstract":"<div><div>In process industries, the complexity and variability of working conditions make it challenging to accurately measure product quality. While data-driven models have developed rapidly, they often overlook the underlying physical or chemical mechanisms. To address this, we propose a hybrid modeling approach that combines mechanism- and data-driven methods. Historical and current working condition data are processed through a hidden layer to extract features. The partial differential equation is discretized and approximated using the forward Euler method to derive mechanism-based quality variable values. These values are then combined with real data through a weighted mix to create a new label for dynamic regression. Additionally, a domain adaptation regularization term is introduced to align the distributions of different working conditions. Through analyses of three process industry datasets, we demonstrate that this method can predict unmeasurable variables with reasonable accuracy and exhibits stronger generalization ability compared to pure data-driven models.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"260 ","pages":"Article 105353"},"PeriodicalIF":3.7,"publicationDate":"2025-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143507522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-18DOI: 10.1016/j.chemolab.2025.105354
Hilthon A. Ramos , Igor Eduardo Silva Arruda , Lucas José de Alencar Danda , Rafaella F. Sales , Julia M. Fernandes , Monica Felts de La Roca Soares , Jose M. Amigo , M. Fernanda Pimentel , José Lamartine Soares Sobrinho
This study explores the potential of spectroscopic analysis combined with Partial Least Squares Regression (PLS) and Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) for the simultaneous quantification of antibiotics in multicomponent drug formulations, specifically clofazimine (CLZ) and dapsone (DAP). The analysis also evaluated the in vitro release profile of the drugs in a fixed-dose combination tablet. High-Performance Liquid Chromatography with Photodiode Array Detection (HPLC-PDA) was used as a reference analytical technique to validate and compare the chemometric models. Both PLS and MCR-ALS models demonstrated high accuracy, with MCR-ALS showing superior predictive capability for CLZ, while both models presented similar performance for DAP quantification. Notably, the results from both models were consistent with the dissolution profile, indicating no statistically significant differences between the spectroscopic and chromatographic quantification methods. Furthermore, the dissolution profile confirmed the immediate release of both active pharmaceutical ingredients (APIs), with no statistically significant differences between the spectroscopic and chromatographic quantification methods. This study highlights the efficiency and versatility of chemometric techniques as an alternative to conventional methods in the quality assessment of anti-leprosy medications.
{"title":"Quantification of antibiotics in multicomponent drug formulations using UV–Vis spectrometer with PLS and MCR-ALS","authors":"Hilthon A. Ramos , Igor Eduardo Silva Arruda , Lucas José de Alencar Danda , Rafaella F. Sales , Julia M. Fernandes , Monica Felts de La Roca Soares , Jose M. Amigo , M. Fernanda Pimentel , José Lamartine Soares Sobrinho","doi":"10.1016/j.chemolab.2025.105354","DOIUrl":"10.1016/j.chemolab.2025.105354","url":null,"abstract":"<div><div>This study explores the potential of spectroscopic analysis combined with Partial Least Squares Regression (PLS) and Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) for the simultaneous quantification of antibiotics in multicomponent drug formulations, specifically clofazimine (CLZ) and dapsone (DAP). The analysis also evaluated the in vitro release profile of the drugs in a fixed-dose combination tablet. High-Performance Liquid Chromatography with Photodiode Array Detection (HPLC-PDA) was used as a reference analytical technique to validate and compare the chemometric models. Both PLS and MCR-ALS models demonstrated high accuracy, with MCR-ALS showing superior predictive capability for CLZ, while both models presented similar performance for DAP quantification. Notably, the results from both models were consistent with the dissolution profile, indicating no statistically significant differences between the spectroscopic and chromatographic quantification methods. Furthermore, the dissolution profile confirmed the immediate release of both active pharmaceutical ingredients (APIs), with no statistically significant differences between the spectroscopic and chromatographic quantification methods. This study highlights the efficiency and versatility of chemometric techniques as an alternative to conventional methods in the quality assessment of anti-leprosy medications.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"260 ","pages":"Article 105354"},"PeriodicalIF":3.7,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474506","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-15DOI: 10.1016/j.chemolab.2025.105352
Mehmet Zahid Malaslı , Mehmet Cabir Akkoyunlu , Engin Pekel , Muhammed Taşova , Samet Kaya Dursun , Mustafa Tahir Akkoyunlu
In the literature have focused on modeling data obtained under drying conditions with different methods and comparing them with each other. However, any studies have been found on estimating the behavior of the same material under different drying conditions. Therefore, a study was conducted to predict the behavior of the same material under different drying conditions. In the study, primarily purple carrot slices were reduced from 6.13 ± 0.05 to 0.14 ± 0.018 g moisture/g dry matter value. Among the models, the drying rates were best estimated by the Midilli-Küçük (R2: 0.9993) model. The lowest energy consumption was determined as 0.285 kWh in the drying process at 70 °C. Estimation of intermediate values is very useful because experimental studies can be length and expensive. Sometimes, even if cost is not a concern, long-term experimental studies and the high number of experiment repetitions increase the importance of estimation methods for researchers. The decision tree, random forest and ada boost methods, which are fast operating methods, were used as estimation methods in this study. MAPE and R2 success values are expressed for all three methods. The Decision Tree method was found to be the most successful technique with the highest R2 value (0.96) and the lowest MAPE value (0.03).
{"title":"Prediction of drying kinetics and energy consumption values of purple carrots dried in a temperature-controlled microwave dryer by decision tree, random forest and ada boost approaches","authors":"Mehmet Zahid Malaslı , Mehmet Cabir Akkoyunlu , Engin Pekel , Muhammed Taşova , Samet Kaya Dursun , Mustafa Tahir Akkoyunlu","doi":"10.1016/j.chemolab.2025.105352","DOIUrl":"10.1016/j.chemolab.2025.105352","url":null,"abstract":"<div><div>In the literature have focused on modeling data obtained under drying conditions with different methods and comparing them with each other. However, any studies have been found on estimating the behavior of the same material under different drying conditions. Therefore, a study was conducted to predict the behavior of the same material under different drying conditions. In the study, primarily purple carrot slices were reduced from 6.13 ± 0.05 to 0.14 ± 0.018 g moisture/g dry matter value. Among the models, the drying rates were best estimated by the Midilli-Küçük (R<sup>2</sup>: 0.9993) model. The lowest energy consumption was determined as 0.285 kWh in the drying process at 70 °C. Estimation of intermediate values is very useful because experimental studies can be length and expensive. Sometimes, even if cost is not a concern, long-term experimental studies and the high number of experiment repetitions increase the importance of estimation methods for researchers. The decision tree, random forest and ada boost methods, which are fast operating methods, were used as estimation methods in this study. <em>MAPE</em> and <em>R</em><sup><em>2</em></sup> success values are expressed for all three methods. The Decision Tree method was found to be the most successful technique with the highest <em>R</em><sup><em>2</em></sup> value (0.96) and the lowest <em>MAPE</em> value (0.03).</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"260 ","pages":"Article 105352"},"PeriodicalIF":3.7,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143527522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-15DOI: 10.1016/j.chemolab.2025.105351
Faramarz Jalili , Ali R. Jalalvand
In this work, a novel chemometrics-assisted electrochemical approach has been developed based on fabrication of a novel electrochemical sensor under computerized methods for simultaneous determination of levodopa (LD), carbidopa (CD) and benserazide (BA) in the presence of indigo carmine (IC) as uncalibrated interference. A glassy carbon electrode (GCE) was modified with multiwalled carbon nanotubes-1-butyl-3-methylimidazolium chloride, [bmim]Cl (MWCNTs-IL), and triple templates molecularly imprinted polymers (TTMIPs) were electrochemically synthesized onto its surface. The effects of experimental parameters on response of the sensor were screened and optimized by Min Run screening and central composite design, respectively. Under optimized conditions, the third-order hydrodynamic differential pulse voltammetric data were recorded and modeled by MCR-ALS, PARAFAC2, U-PLS/RTL, N-PLS-RTL, U-PCA/RTL, and APARAFAC to select the best algorithm for assisting the sensor with the aim of simultaneous determination of LD, CD and BA in the presence of IC as uncalibrated interference. Our results confirmed MCR-ALS showed the best performance to assist the sensor for the analysis of synthetic samples. The TTMIPs/MWCNTs-IL/GCE assisted by MCR-ALS was also successful in analysis of pharmaceuticals used as medications to the treatment of Parkinson's disease, and its performance was comparable with HPLC-UV as the refence method.
{"title":"Developing a novel and intelligent chemometrics-assisted molecularly imprinted electrochemical sensor: Application to the improvement of the efficiency of the treatment of Parkinson's disease","authors":"Faramarz Jalili , Ali R. Jalalvand","doi":"10.1016/j.chemolab.2025.105351","DOIUrl":"10.1016/j.chemolab.2025.105351","url":null,"abstract":"<div><div>In this work, a novel chemometrics-assisted electrochemical approach has been developed based on fabrication of a novel electrochemical sensor under computerized methods for simultaneous determination of levodopa (LD), carbidopa (CD) and benserazide (BA) in the presence of indigo carmine (IC) as uncalibrated interference. A glassy carbon electrode (GCE) was modified with multiwalled carbon nanotubes-1-butyl-3-methylimidazolium chloride, [bmim]Cl (MWCNTs-IL), and triple templates molecularly imprinted polymers (TTMIPs) were electrochemically synthesized onto its surface. The effects of experimental parameters on response of the sensor were screened and optimized by Min Run screening and central composite design, respectively. Under optimized conditions, the third-order hydrodynamic differential pulse voltammetric data were recorded and modeled by MCR-ALS, PARAFAC2, U-PLS/RTL, N-PLS-RTL, U-PCA/RTL, and APARAFAC to select the best algorithm for assisting the sensor with the aim of simultaneous determination of LD, CD and BA in the presence of IC as uncalibrated interference. Our results confirmed MCR-ALS showed the best performance to assist the sensor for the analysis of synthetic samples. The TTMIPs/MWCNTs-IL/GCE assisted by MCR-ALS was also successful in analysis of pharmaceuticals used as medications to the treatment of Parkinson's disease, and its performance was comparable with HPLC-UV as the refence method.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"259 ","pages":"Article 105351"},"PeriodicalIF":3.7,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429894","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-14DOI: 10.1016/j.chemolab.2025.105340
Christian Ickes , Pirya Rani , Kristiyana Tsenova , Johanna Echternach , Frank Führer , Detlef Bartel , Christel Kamp
Raman spectroscopy is a widely used technique for the identification of chemical substances and in the quality control of pharmaceutical products. Inelastic scattering of laser light generates unique fingerprints of chemical substances which allows for identification of products and quantification of active components. Using this non-destructive technique for biomedicines like vaccines or therapeutic allergen products introduces new challenges in terms of experimental setup, spectral processing, and their standardization. We explore experimental setups and use machine learning techniques to evaluate the potential of Raman spectroscopy to distinguish between therapeutic allergen products from different manufacturers with closely related bee and wasp venoms as Active Pharmaceutical Ingredients (APIs). A comparison of various models shows that a differentiation of products is possible based on their Raman spectra at accuracies above 95%. A deeper analysis allows to identify key regions in the spectra for differentiation. These can guide further research towards the identification and quantification of biochemical compounds of interest. In conclusion, this proof-of-concept study shows the applicability of Raman spectroscopy in the quality assurance of biomedicines and suggests directions for further in-depth analyses.
{"title":"Identification of therapeutic allergen products using their Raman spectral fingerprint","authors":"Christian Ickes , Pirya Rani , Kristiyana Tsenova , Johanna Echternach , Frank Führer , Detlef Bartel , Christel Kamp","doi":"10.1016/j.chemolab.2025.105340","DOIUrl":"10.1016/j.chemolab.2025.105340","url":null,"abstract":"<div><div>Raman spectroscopy is a widely used technique for the identification of chemical substances and in the quality control of pharmaceutical products. Inelastic scattering of laser light generates unique fingerprints of chemical substances which allows for identification of products and quantification of active components. Using this non-destructive technique for biomedicines like vaccines or therapeutic allergen products introduces new challenges in terms of experimental setup, spectral processing, and their standardization. We explore experimental setups and use machine learning techniques to evaluate the potential of Raman spectroscopy to distinguish between therapeutic allergen products from different manufacturers with closely related bee and wasp venoms as Active Pharmaceutical Ingredients (APIs). A comparison of various models shows that a differentiation of products is possible based on their Raman spectra at accuracies above 95%. A deeper analysis allows to identify key regions in the spectra for differentiation. These can guide further research towards the identification and quantification of biochemical compounds of interest. In conclusion, this proof-of-concept study shows the applicability of Raman spectroscopy in the quality assurance of biomedicines and suggests directions for further in-depth analyses.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"260 ","pages":"Article 105340"},"PeriodicalIF":3.7,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143507521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-13DOI: 10.1016/j.chemolab.2025.105337
Bihter Das, Huseyin Alperen Dagdogen, Muhammed Onur Kaya, Ozkan Tuncel, Muhammed Samet Akgul, Resul Das
Drug–drug interactions (DDIs) are a significant challenge in modern healthcare, especially in polypharmacy, where patients are given more than one drug at the same time. Accurate prediction of DDIs plays an important role in reducing adverse effects and improving recovery in patients. In this study, we propose GAINET, a derivative of the graph-based neural network model enhanced with attention mechanisms, to accurately improve the prediction of drug–drug interactions. The model effectively learns interaction models by focusing on critical features in drug structures and their interactions with each other through molecular graph representations. For the performance evaluation of GAINET, which is trained on the DrugBank dataset containing 191,870 DDI examples, basic metrics such as AUC-ROC, F1 score, precision and recall are used. The obtained accuracy of 0.9050, F1 score of 0.9096 and AUC-ROC of 0.9505 indicate that GAINET outperforms many state-of-the-art models and has good generalization ability even on previously untested data. Moreover, the molecular attention mechanism enables interpretable predictions by highlighting the interaction-specific molecular substructures. All these findings indicate that GAINET, our proposed model for DDI prediction, can serve as a valuable and useful tool and advance the development of reliable pharmacological treatments.
{"title":"GAINET: Enhancing drug–drug interaction predictions through graph neural networks and attention mechanisms","authors":"Bihter Das, Huseyin Alperen Dagdogen, Muhammed Onur Kaya, Ozkan Tuncel, Muhammed Samet Akgul, Resul Das","doi":"10.1016/j.chemolab.2025.105337","DOIUrl":"10.1016/j.chemolab.2025.105337","url":null,"abstract":"<div><div>Drug–drug interactions (DDIs) are a significant challenge in modern healthcare, especially in polypharmacy, where patients are given more than one drug at the same time. Accurate prediction of DDIs plays an important role in reducing adverse effects and improving recovery in patients. In this study, we propose GAINET, a derivative of the graph-based neural network model enhanced with attention mechanisms, to accurately improve the prediction of drug–drug interactions. The model effectively learns interaction models by focusing on critical features in drug structures and their interactions with each other through molecular graph representations. For the performance evaluation of GAINET, which is trained on the DrugBank dataset containing 191,870 DDI examples, basic metrics such as AUC-ROC, F1 score, precision and recall are used. The obtained accuracy of 0.9050, F1 score of 0.9096 and AUC-ROC of 0.9505 indicate that GAINET outperforms many state-of-the-art models and has good generalization ability even on previously untested data. Moreover, the molecular attention mechanism enables interpretable predictions by highlighting the interaction-specific molecular substructures. All these findings indicate that GAINET, our proposed model for DDI prediction, can serve as a valuable and useful tool and advance the development of reliable pharmacological treatments.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"259 ","pages":"Article 105337"},"PeriodicalIF":3.7,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421365","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In industrial environments, the unpredictability and irreproducibility of faults often result in insufficient sample sizes and atypical data features, significantly increasing the challenges faced by traditional fault diagnosis methods. To address these issues, this paper proposes a novel fault diagnosis approach that integrates the Borderline embedded deep synthetic minority oversampling technique (BE-DeepSMOTE) with Laplacian matrix decomposition, with the aim of tackling fault identification problems in imbalanced data scenarios. BE-DeepSMOTE employs a deep encoder–decoder framework to enable end-to-end learning and reconstruction of multi-dimensional features. It further incorporates the Borderline SMOTE technique to oversample minority class instances in the feature space, thereby enhancing their representation while ensuring statistical consistency with the original dataset to mitigate data imbalance. Furthermore, we introduce an ensemble classifier that combines Adaboost with Laplacian matrix decomposition. This ensemble classifier leverages the synergy of multiple weak classifiers to extract geometric properties and graph structure similarities from the data, while employing an adaptive weighting mechanism to improve the diagnostic accuracy. Experimental results from two industrial processes demonstrate that the proposed approach significantly enhances the diagnostic accuracy and stability in imbalanced instance environments.
{"title":"DeepSMOTE with Laplacian matrix decomposition for imbalance instance fault diagnosis","authors":"Yuan Xu, Rui-Ze Fan, Yan-Lin He, Qun-Xiong Zhu, Yang Zhang, Ming-Qing Zhang","doi":"10.1016/j.chemolab.2025.105338","DOIUrl":"10.1016/j.chemolab.2025.105338","url":null,"abstract":"<div><div>In industrial environments, the unpredictability and irreproducibility of faults often result in insufficient sample sizes and atypical data features, significantly increasing the challenges faced by traditional fault diagnosis methods. To address these issues, this paper proposes a novel fault diagnosis approach that integrates the Borderline embedded deep synthetic minority oversampling technique (BE-DeepSMOTE) with Laplacian matrix decomposition, with the aim of tackling fault identification problems in imbalanced data scenarios. BE-DeepSMOTE employs a deep encoder–decoder framework to enable end-to-end learning and reconstruction of multi-dimensional features. It further incorporates the Borderline SMOTE technique to oversample minority class instances in the feature space, thereby enhancing their representation while ensuring statistical consistency with the original dataset to mitigate data imbalance. Furthermore, we introduce an ensemble classifier that combines Adaboost with Laplacian matrix decomposition. This ensemble classifier leverages the synergy of multiple weak classifiers to extract geometric properties and graph structure similarities from the data, while employing an adaptive weighting mechanism to improve the diagnostic accuracy. Experimental results from two industrial processes demonstrate that the proposed approach significantly enhances the diagnostic accuracy and stability in imbalanced instance environments.</div></div>","PeriodicalId":9774,"journal":{"name":"Chemometrics and Intelligent Laboratory Systems","volume":"259 ","pages":"Article 105338"},"PeriodicalIF":3.7,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143403647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号