Article
Computer Science, Interdisciplinary Applications
Ahmed H. Alkenani, Yuefeng Li, Yue Xu, Qing Zhang
Summary: The study emphasizes the importance of automating the diagnosis of AD using language deficiency, developing multiple heterogeneous stacked fusion models to improve generalizability and robustness of AD diagnostic ML models. The models trained on two different datasets achieved high AUC, accuracy, and F1 score values. The suggestion is to replace traditional screening tests with these models for fully automated remote diagnosis.
JOURNAL OF BIOMEDICAL INFORMATICS
(2021)
Article
Computer Science, Information Systems
Geonu Lee, Kimin Yun, Jungchan Cho
Summary: This study introduces a new framework OSGNet for HOI detection, which solves the issues with CNN-based HOI detection and exhibits good generalization performance. The proposed method achieves state-of-the-art accuracy through training sub-models and meta-learner, demonstrating excellent performance on unseen test data.
Article
Chemistry, Analytical
Sunil Kumar Prabhakar, Harikumar Rajaguru, Semin Ryu, In Cheol Jeong, Dong-Ok Won
Summary: Manual sleep stage scoring is a hectic task, leading to the development of automated sleep stage classification systems. This study proposes a holistic strategy combining clustering, dimensionality reduction, feature extraction and selection, and deep learning for sleep stage classification. The methodology surpasses previous studies in terms of classification accuracy, reporting a high accuracy of 93.51% even for a six-class classification problem.
Article
Computer Science, Artificial Intelligence
K. Aditya Shastry, H. A. Sanjay
Summary: Data pre-processing is a technique that transforms raw data into a useful format for machine learning, with feature selection and feature extraction being significant components. This study proposes a hybrid strategy using modified Genetic Algorithm and weighted Principal Component Analysis for selecting and extracting features from agricultural datasets, resulting in significant improvements in benchmark and real-world farming datasets.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Microbiology
Zahoor Ahmed, Hasan Zulfiqar, Abdullah Aman Khan, Ijaz Gul, Fu-Ying Dao, Zhao-Yue Zhang, Xiao-Long Yu, Lixia Tang
Summary: Thermophilic proteins have important application value in biotechnology and industrial processes. A multi-layer perceptron (MLP) model based on a multi-feature fusion strategy was proposed to accurately identify thermophilic proteins. The developed software package iThermo provides a user-friendly platform for conveniently applying the model.
FRONTIERS IN MICROBIOLOGY
(2022)
Article
Automation & Control Systems
Victor Hamer, Pierre Dupont
Summary: Current feature selection methods, especially in high-dimensional data, may suffer from instability, but a new stability measure proposed in this work, which incorporates the importance of selected features in predictive models, has been shown to correct overly optimistic estimates and improve decision-making accuracy.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Computer Science, Information Systems
Harshit Kaushik, Dilbag Singh, Manjit Kaur, Hammam Alshazly, Atef Zaguia, Habib Hamam
Summary: The research proposes a new method to eliminate unnecessary reflectance properties of eye fundus images using image processing and deep learning techniques, aiming to improve the diagnostic analysis of diabetic retinopathy.
Article
Engineering, Mechanical
Siqi Shi, Shijie Jin, Donghui Zhang, Jingyu Liao, Dongxin Fu, Li Lin
Summary: This paper proposes a generally applicable machine learning framework based on the model interpretation strategy to improve the accuracy and efficiency of ultrasonic testing. The framework utilizes signal processing techniques to extract features and integrates multiple feature selection methods to determine the optimal feature subset and machine learning model adaptively. Experimental results demonstrate the effectiveness of the framework in identifying and locating side-drilled holes.
CHINESE JOURNAL OF MECHANICAL ENGINEERING
(2023)
Article
Chemistry, Analytical
Oussama Djedidi, Mohand A. Djeziri, Nicolas Morati, Jean-Luc Seguin, Marc Bendahan, Thierry Contaret
Summary: The approach described in this work utilizes a single physical sensor and data-driven algorithms to detect the presence of the dangerous gases CO, NO2, and O3 individually or in mixtures. By ranking and selecting the best features, a multi-Support Vector Machine model is trained and validated to further enhance classification results, showcasing the effectiveness of the proposed approach in gas detection and discrimination.
SENSORS AND ACTUATORS B-CHEMICAL
(2021)
Article
Biochemical Research Methods
Fengsheng Wang, Leyi Wei
Summary: In this study, we propose a novel multi-scale end-to-end deep learning model, MSTLoc, for identifying protein subcellular locations in the imbalanced multi-label immunohistochemistry (IHC) images dataset. We demonstrate that the proposed MSTLoc outperforms current state-of-the-art models in multi-label subcellular location prediction. Through feature visualization and interpretation analysis, we show that the multi-scale deep features learned from our model exhibit better ability in capturing discriminative patterns underlying protein subcellular locations, and the features from different scales are complementary for the improvement in performance. Case study results indicate that our MSTLoc can successfully identify some biomarkers from proteins that are closely involved in cancer development.
Article
Construction & Building Technology
Behzad Najafi, Monica Depalo, Fabio Rinaldi, Reza Arghandeh
Summary: The study focuses on extracting influential features from smart meter data to improve machine learning-based classification of non-residential buildings. Through advanced feature selection methods and a custom approach, the number of features needed for classification is reduced while accuracy is increased. By selecting and utilizing fewer features, the methodology simplifies feature extraction procedures and enhances interpretation of important features' influence.
ENERGY AND BUILDINGS
(2021)
Article
Computer Science, Information Systems
Hua Chen, Kehui Mei, Yuan Zhou, Nan Wang, Guangxing Cai
Summary: This study focuses on breast cancer and proposes a hybrid strategy combined with machine learning methods to build an accurate and efficient breast cancer auxiliary diagnosis model. Experimental results show that the new approach achieves better prediction results compared to previous methods.
Article
Automation & Control Systems
Adel Afia, Fawzi Gougam, Chemseddine Rahmoune, Walid Touzout, Hand Ouelmokhtar, Djamel Benazzouz
Summary: This paper develops an intelligent algorithm for diagnosing faults in reciprocating air compressors using real-time acoustic signals. The algorithm consists of three steps: feature extraction, selection, and classification. Experimental acoustic signals are decomposed using maximal overlap discrete wavelet packet transform and time domain features are calculated to build health state feature matrices. Features are selected using Harris hawks optimization, and then classified using random forest, ensemble tree, and K-nearest neighbors algorithms. Comparative studies prove the efficiency of the proposed approach in fault detection and classification of air compressors.
TRANSACTIONS OF THE INSTITUTE OF MEASUREMENT AND CONTROL
(2023)
Article
Engineering, Chemical
Zhiqiang Wang, Xintong Zhang, Dakuo He
Summary: In this paper, a dynamic global feature extraction (DGFE) method combining principal component analysis (PCA) and kernel principal component analysis (KPCA) is proposed to mine the dynamic characteristics of large-scale industrial data. Additionally, a new importance-correlation-based feature selection (ICFS) method is introduced to ensure the optimality of the obtained feature set. Experimental results on a copper flotation industrial process demonstrate the effectiveness of the proposed methods.
CANADIAN JOURNAL OF CHEMICAL ENGINEERING
(2023)
Article
Environmental Sciences
Yaxuan Huang, Bin Guo, Haoxuan Sun, Huijie Liu, Song Xi Chen
Summary: The study found regularities in the meteorological processes of air pollution in six major cities in North China, with PM2.5, PM10, SO2, and CO being mainly influenced by dew point temperature and air pressure, and NO2 and O3 being mostly impacted by boundary layer height and temperature. The research also showed that boundary layer height could be accurately modeled using surface meteorological variables, indicating that air quality assessment without considering BLH would still yield satisfactory results.
ATMOSPHERIC ENVIRONMENT
(2021)
Article
Environmental Sciences
Toshimi Nakajima, Mao Kuragano, Makoto Yamada, Ryo Sugimoto
Summary: This study compared the contribution of submarine groundwater discharge (SGD) to river nutrient budgets at nearshore and embayment scales, and found that SGD-derived nutrients become more important at larger spatial scales.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Fan Liu, Lei Zhang, Chongyang Zhang, Ziguang Chen, Jingguang Li
Summary: NO2 emissions from wall-mounted gas stoves used for household heating have become a significant source of indoor pollution in Chinese urban areas. The high indoor concentration of NO2 poses potential health risks to residents. It is urgently necessary to establish relevant regulations and implement emission reduction technologies to reduce NO2 emissions from wall-mounted gas stoves.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Letter
Environmental Sciences
Hans Peter H. Arp, Raoul Wolf, Sarah E. Hale, Sivani Baskaran, Juliane Gluege, Martin Scheringer, Xenia Trier, Ian T. Cousins, Harrie Timmer, Roberta Hofman-Caris, Anna Lennquist, Andre D. Bannink, Gerard J. Stroomberg, Rosa M. A. Sjerps, Rosa Montes, Rosario Rodil, Jose Benito Quintana, Daniel Zahn, Herve Gallard, Tobias Mohr, Ivo Schliebner, Michael Neumann
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Philomina Onyedikachi Peter, Binessi Edouard Ifon, Francois Nkinahamira, Kayode Hassan Lasisi, Jiangwei Li, Anyi Hu, Chang-Ping Yu
Summary: This study investigates the relationship between dissolved organic matter (DOM) and Rare Earth Elements (REEs) in sediments from Yundang Lagoon, China. The results show four distinct fluorescent components, with protein-like substances being the most prevalent. Additionally, the total fluorescence intensity and LREE concentrations exhibit a synchronized increase from Outer to Inner to Songbai Lake core sediments. The findings demonstrate a strong correlation between DOM content and pollution levels.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Surya Gupta, Pasquale Borrelli, Panos Panagos, Christine Alewell
Summary: The objective of this study is to incorporate soil hydraulic properties into the erodibility factor (K) of USLE-type models. By modifying and improving the existing equations for soil texture and permeability, the study successfully included information on saturated hydraulic conductivity (Ksat) into the calculation of K factor. Using the Random Forest machine learning algorithm, two independent K factor maps with different spatial resolutions were generated. The results show that the decrease in K factor values has a positive impact on the modeling of soil erosion rates.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Jesmin Akter, Wendy J. M. Smith, Yawen Liu, Ilho Kim, Stuart L. Simpson, Phong Thai, Asja Korajkic, Warish Ahmed
Summary: The choice of workflow in wastewater surveillance has a significant impact on SARS-CoV-2 concentrations, while having minimal effects on HF183 and no effect on HAdV 40/41 concentrations. Certain components in the workflow can be interchangeable, but factors such as buffer type, chloroform, and homogenization speed can affect the recovery of viruses and bacteria.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Yu Luo, Xueting Yang, Diwei Wang, Hongmei Xu, Hongai Zhang, Shasha Huang, Qiyuan Wang, Ningning Zhang, Junji Cao, Zhenxing Shen
Summary: Atmospheric PM2.5, which can generate reactive oxygen species (ROS), is associated with cardiorespiratory morbidity and mortality. The study found that both the mass concentration of PM2.5 and the DTT activity were higher during the heating season than during the nonheating season. Combustion sources were the primary contributors to DTT activity during the heating season, while secondary formation dominated during the nonheating season. The study also revealed that biomass burning had the highest inherent oxidation potential among all sources investigated.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Erin L. Murphy, Leah R. Gerber, Chelsea M. Rochman, Beth Polidoro
Summary: Plastic pollution has devastating consequences for marine organisms. This study uses a trait-based framework to develop a vulnerability index for marine mammals, seabirds, and sea turtles in Hawai'i. The index ranks 63 study species based on their vulnerability to macroplastic pollution, providing valuable information for species monitoring and management priorities.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Kenji Maurice, Amelia Bourceret, Sami Youssef, Stephane Boivin, Liam Laurent-Webb, Coraline Damasio, Hassan Boukcim, Marc-Andre Selosse, Marc Ducousso
Summary: Growing pressure from climate change and agricultural land use is destabilizing soil microbial community interactions. Little is known about microbial community resistance and adaptation to disturbances, hindering our understanding of recovery latency and implications for ecosystem functioning. This study found that anthropic disturbance and natural disturbance have different effects on the topology and stability of soil microbial networks.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Yunhao Li, Yali Feng, Haoran Li, Yisong Yao, Chenglong Xu, Jinrong Ju, Ruiyu Ma, Haoyu Wang, Shiwei Jiang
Summary: Deep-sea mining poses a serious threat to marine ecosystems and human health by disturbing sediment and transmitting metal ions through the food chain. This study developed a new regenerative adsorption material, OMN@SA, which effectively removes metal ions. The adsorption mechanism and performance of the material for metal ion fixation were investigated.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Antonio Medici, Margherita Lavorgna, Marina Isidori, Chiara Russo, Elena Orlo, Giovanni Luongo, Giovanni Di Fabio, Armando Zarrelli
Summary: Valsartan, a widely used antihypertensive drug, has been detected in high concentrations in surface waters due to its unchanged excretion and incomplete degradation in wastewater treatment plants. This study investigated the degradation of valsartan and identified 14 degradation byproducts. The acute and chronic toxicity of these byproducts were evaluated in key organisms in the freshwater trophic chain.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Jiang Lin, Lianbao Chi, Qing Yuan, Busu Li, Mingbao Feng
Summary: This study investigated the photodegradation behavior and product formation of two representative pharmaceuticals in simulated estuary water. The study found that the formed transformation products of these pharmaceuticals have potential toxicity on marine organisms, including oxidative stress and damage to cellular components.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Hua Fang, Dongdong Jiang, Ye He, Siyi Wu, Yuehong Li, Ziqi Zhang, Haoting Chen, Zixin Zheng, Yan Sun, Wenxiang Wang
Summary: This study revealed that exposure to lower levels of air pollutants led to decreased pregnancy rates, with PM10, NO2, SO2, and CO emerging as the four most prominent pollutants. Individuals aged 35 and above exhibited heightened susceptibility to pollutants.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Ali Shaan Manzoor Ghumman, Rashid Shamsuddin, Amin Abbasi, Mohaira Ahmad, Yoshiaki Yoshida, Abdul Sami, Hamad Almohamadi
Summary: In this study, inverse vulcanized polysulfides (IVP) were synthesized by reacting molten sulfur with 4-vinyl benzyl chloride, and then functionalized using N-methyl D-glucamine (NMDG). The functionalized IVP showed a high mercury adsorption capacity and a machine learning model was developed to predict the amount of mercury removed. Furthermore, the functionalized IVP can be regenerated and reused, providing a sustainable and cost-effective adsorbent.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)
Article
Environmental Sciences
Rita Bonfiglio, Renata Sisto, Stefano Casciardi, Valeria Palumbo, Maria Paola Scioli, Erica Giacobbi, Francesca Servadei, Gerry Melino, Alessandro Mauriello, Manuel Scimeca
Summary: This study investigated the presence of aluminum in human colon cancer samples and its potential association with biological processes involved in cancer progression. Aluminum was found in tumor areas of 24% of patients and was associated with epithelial to mesenchymal transition (EMT) and cell death. Additional analyses revealed higher tumor mutational burden and mutations in genes related to EMT and apoptosis in aluminum-positive colon cancers. Understanding the molecular mechanisms of aluminum toxicity may improve strategies for the management of colon cancer patients.
SCIENCE OF THE TOTAL ENVIRONMENT
(2024)