Article
Computer Science, Information Systems
Debasmita GhoshRoy, Parvez Ahmad Alvi, K. C. Santosh
Summary: Infertility is a global issue, with male factors contributing to about 40% to 50% of cases. Existing AI systems lack interpretability, limiting clinicians' understanding of decision-making processes and their application in healthcare. This study introduces an explainable model for male fertility prediction, utilizing nine features related to lifestyle and environmental factors. Five AI tools (support vector machine, adaptive boosting, XGB, random forest, and extra tree algorithms) are deployed with balanced and imbalanced datasets, incorporating explainable AI techniques such as LIME, SHAP, and ELI5. XGB outperformed existing AI systems, achieving an optimal AUC of 0.98.
Article
Engineering, Civil
Deva Charan Jarajapu, Maheswaran Rathinasamy, Ankit Agarwal, Axel Bronstert
Summary: Regional Flood Frequency Analysis (RFFA) is a widely used approach for estimating design floods. This study developed an XGB-based machine learning model for RFFA and flood estimation, which showed high accuracy in estimating design flood and visualized the importance of catchment features.
JOURNAL OF HYDROLOGY
(2022)
Article
Engineering, Civil
Hoang-Anh Le, Duc-Anh Le, Thanh-Tung Le, Hoai-Phuong Le, Thanh-Hai Le, Huong-Giang Thi Hoang, Thuy-Anh Nguyen
Summary: This study develops a reliable and accurate machine learning (ML) model to predict the shear strength of fiber-reinforced polymer (FRP) reinforced concrete (RC) beams. Through feature importance analysis and data techniques, the most significant features affecting the shear strength of the beams are identified.
Article
Computer Science, Interdisciplinary Applications
Sajjad Mirzaei, Mehdi Vafakhah, Biswajeet Pradhan, Seyed Jalil Alavi
Summary: This study implemented extreme gradient boosting (EGB) method for flood susceptibility modelling and compared its performance with three advanced benchmark models. Results showed that Random Forest (RF) model and Extreme Gradient Boosting had the best performance. The study found that factors such as distance from rivers have important influence on flood susceptibility mapping, and recommended the application of RF and EGB models for such studies.
EARTH SCIENCE INFORMATICS
(2021)
Article
Thermodynamics
Ahmet Hasim Yurttakal
Summary: The study developed a model for estimating thermal conductivity values using the extreme gradient boosting algorithm, and evaluated the performance of the model on unseen test data. The algorithm achieved encouraging results with 0.18 RAISE, 0.99 R-2, and 3.18% MAE values.
Article
Environmental Sciences
Hossein Sahour, Vahid Gholami, Javad Torkaman, Mehdi Vazifedan, Sirwe Saeedi
Summary: Monitoring temporal variation of streamflow is crucial for water resources management, and a machine learning framework utilizing tree-rings and vessel features was proposed in this study. Results indicate that the extreme gradient boosting model outperforms the random forest model in streamflow modeling, particularly for normal and low flows.
ENVIRONMENTAL EARTH SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Jiaming Han, Kunxin Shu, Zhenyu Wang
Summary: Annual increases in global energy consumption are inevitable due to a growing global economy and population. Among sectors, the construction industry consumes 20.1% of the world's total energy, making it crucial to explore methods for estimating energy usage. Various computational approaches exist, including statistics-based, engineering-based, and machine learning-based methods. Machine learning-based frameworks outperform the others. In our study, we propose using the Extreme Gradient Boosting algorithm to predict energy consumption, achieving better results with combined historical and date features.
PEERJ COMPUTER SCIENCE
(2023)
Article
Engineering, Manufacturing
Sarbari Ganguly, Sougat Manna
Summary: Continuous cooling transformation (CCT) diagram is a crucial tool in the steel industry for new product development and is traditionally obtained through extensive thermomechanical studies. In this work, a machine-learning model using Light Gradient Boosting Machine (LGBM) was developed to predict the CCT of a new steel grade, considering its chemistry and processing parameters. The model was validated against test data and well-established metallurgical correlations, demonstrating good performance.
MATERIALS AND MANUFACTURING PROCESSES
(2023)
Article
Thermodynamics
Mohamed Massaoudi, Shady S. Refaat, Ines Chihi, Mohamed Trabelsi, Fakhreddine S. Oueslati, Haitham Abu-Rub
Summary: This paper introduces an effective computing framework for Short-Term Load Forecasting (STLF) using stacked generalization approach with three models, demonstrating its performance through validation and contributions in novel algorithm, effective technique, critical analysis for hyperparameter optimization, and comparative study.
Article
Environmental Sciences
Chao-Yu Guo, Ke-Hao Chang
Summary: This article explores the importance of interaction effects in cardiac research and proposes a new machine learning algorithm for assessing the p-value of feature interaction. The algorithm outperforms traditional statistical models in simulated studies.
INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH
(2022)
Article
Computer Science, Information Systems
J. Jasmine Gabriel, L. Jani Anbarasi
Summary: Coronary Artery Disease (CAD) is a prevalent disorder that requires low-cost automated technology for early diagnosis and treatment. Traditional machine learning methods may not be suitable for smaller clinical datasets with categorical features, necessitating alternative approaches such as data preprocessing and feature selection. The developed BSOXGB model achieves remarkable accuracy on the Z-Alizadeh Sani dataset, making it a practical solution for automatically detecting and diagnosing CAD.
Article
Materials Science, Multidisciplinary
Seungro Lee, Joonhee Park, Naksoo Kim, Taeyong Lee, Luca Quagliato
Summary: This paper presents a machine learning methodology that can learn from simulation results, experimental data, or sensor signals, and is capable of predicting and optimizing specific user-defined process and design parameters. The methodology utilizes an enhanced Extreme Gradient Boosting (XGB) algorithm and a metaheuristic search algorithm based on Differential Evolution (DE) architecture for optimization.
MATERIALS & DESIGN
(2023)
Article
Environmental Sciences
Meng Wang, Xue Li, Mei Lei, Lunbo Duan, Huichao Chen
Summary: The petrochemical industry is a key industry in soil pollution, with significant effects on human health and the ecological environment. This study developed an efficient method using the XGBoost algorithm for identifying health risks in the petrochemical industry. The results showed that XGBoost outperformed other models in health risk identification.
ECOTOXICOLOGY AND ENVIRONMENTAL SAFETY
(2022)
Article
Thermodynamics
Qichao Lv, Ali Rashidi-Khaniabadi, Rong Zheng, Tongke Zhou, Mohammad-Reza Mohammadi, Abdolhossein Hemmati-Sarapardeh
Summary: In this study, five machine learning models based on Gaussian process regression (GPR) and Extreme gradient boosting (XGBoost) were developed to estimate the diffusion coefficient of CO2 in heavy crude oil/bitumen. The XGBoost model demonstrated the highest precision with an R2 of 0.9998 and an average absolute percent relative error of 0.68%. The trends analysis showed that the diffusion coefficient of CO2 in bitumen is a unimodal function of gas concentration, while temperature and pressure have increasing effects on the CO2 diffusion coefficient, which were accurately predicted by the XGBoost model.
Article
Medicine, General & Internal
N. Casillas, A. M. Torres, M. Moret, A. Gomez, J. M. Rius-Peris, J. Mateo
Summary: Recently, there has been an increased demand for assistance in global health due to the COVID-19 pandemic. Researchers have conducted studies using machine learning techniques to find variables associated with increased clinical risk and effective treatments. This study used the XGBoost algorithm to accurately predict mortality rates in COVID-19 patients.
INTERNAL AND EMERGENCY MEDICINE
(2022)
Article
Management
Ulpan Tokkozhina, Ana Lucia Martins, Joao C. Ferreira
Summary: This study explores the dimensions of impact of blockchain technology adoption in supply chain management and discusses the synergetic and counter-synergetic effects between these dimensions. It identifies three dimensions - 'operations and processes', 'supply chain relationships', and 'innovation and data access' - and demonstrates how they overlap and interact, leading to both positive and negative effects. The study provides valuable insights for scholars and practitioners to prevent undesired effects and enhance desired ones.
OPERATIONS MANAGEMENT RESEARCH
(2023)
Article
Chemistry, Analytical
Stefan Postolache, Pedro Sebastiao, Vitor Viegas, Octavian Postolache, Francisco Cercas
Summary: Soil nutrients assessment is important in horticulture, but implementing an information system for horticulture faces challenges such as spatial variability, different soil properties for different plants, varying nutrient uptake, small size of monoculture, and diversity in farm components and socio-economic factors. However, advances in technology allow for the creation of low-cost, efficient information systems to improve resource management and increase productivity and sustainability in horticultural farms.
Review
Chemistry, Analytical
Zahra Mardani Korani, Armin Moin, Alberto Rodrigues da Silva, Joao Carlos Ferreira
Summary: This paper reviews the literature on model-driven engineering (MDE) tools and languages for the internet of things (IoT) with a focus on data analytics and machine learning techniques. The paper examines prior work in this area and classifies it based on its support for DAML techniques, especially time series analysis. The key research questions addressed in the paper are the proposed MDE approaches, tools, and languages and their support for DAML techniques in the context of smart IoT services.
Article
Chemistry, Analytical
Antonio Raimundo, Joao Pedro Pavia, Pedro Sebastiao, Octavian Postolache
Summary: Industrial inspection is crucial for quality and safety; this paper proposes YOLOX-Ray, an efficient deep learning architecture for industrial inspection; through combining SimAM attention mechanism and Alpha-IoU loss function, YOLOX-Ray outperforms other configurations in three case studies.
Article
Chemistry, Multidisciplinary
Joana Fogaca, Tomas Brandao, Joao C. Ferreira
Summary: This research work aims to develop a system that can automatically detect illegal graffiti in real-time in Lisbon using cars equipped with cameras. A classification model with an overall accuracy of 81.4% is used to classify images into street art, illegal graffiti, or no graffiti. Another model is trained to detect the coordinates of graffiti on an image, achieving an Intersection over Union (IoU) of 70.3% for the test set.
APPLIED SCIENCES-BASEL
(2023)
Article
Management
Ulpan Tokkozhina, Ana Lucia Martins, Joao C. Ferreira
Summary: This study examines the consequences of information availability in the context of blockchain technology (BCT) adoption pilots in a Portuguese multi-tier supply chain for frozen fish products. The study found that trustful relationships between players are still necessary prior to BCT adoption, despite its reputation as a trust-enabling technology. Additionally, the study revealed a higher probability of purchasing fish products that have traceable information available, even when traceability is not a major concern for final consumers.
OPERATIONS MANAGEMENT RESEARCH
(2023)
Article
Computer Science, Information Systems
Joao Monge, Goncalo Ribeiro, Antonio Raimundo, Octavian Postolache, Joel Santos
Summary: Health monitoring is crucial in healthcare facilities, but challenges such as human error, patient compliance concerns, and limited resources affect the reliability of health data. To address these issues, we propose a non-intrusive smart sensing system using SensFloor smart carpet and an IMU wearable sensor to monitor position and gait characteristics. Machine learning algorithms are utilized to analyze the collected data and generate real-time, cloud-stored information accessible to medical professionals and patients. The system's real-time dashboards provide comprehensive analysis, enabling personalized training plans and better rehabilitation outcomes.
Article
Green & Sustainable Science & Technology
Paola Andrea de Antonio Boada, Julian Fernando Ordonez Duran, Fabio Leonardo Gomez Avila, Joao Carlos Espindola Ferreira
Summary: This study evaluates the presence of sustainability parameters in the product development process, finds that sustainability has independence and relevance in product planning, and allows for short, medium, and long-term actions. The characteristics and development potential of two ecosystems are compared, and a sustainable baseline is established.
Article
Health Care Sciences & Services
Luis B. Elvas, Miguel Nunes, Joao C. Ferreira, Miguel Sales Dias, Luis Bras Rosario
Summary: This study utilizes exploratory data analysis and predictive machine learning models based on hospital data to provide rapid and accurate tools for diagnosing and intervening in cardiovascular diseases, which has significant clinical importance.
JOURNAL OF PERSONALIZED MEDICINE
(2023)
Article
Health Care Sciences & Services
Jose Pereira, Nuno Antunes, Joana Rosa, Joao C. Ferreira, Sandra Mogo, Manuel Pereira
Summary: This research reviews recent works in health remote monitoring systems for COPD patients and develops an intelligent clinical decision support system (CIDSS) that provides early health information and risk analysis to support COPD treatment.
JOURNAL OF PERSONALIZED MEDICINE
(2023)
Article
Pediatrics
Joao Rala Cordeiro, Sara Mosca, Ana Correia-Costa, Catia Ferreira, Joana Pimenta, Liane Correia-Costa, Henrique Barros, Octavian Postolache
Summary: This study explores the impact of overweight and obesity during childhood on cardiovascular health using data science techniques. The findings suggest that obesity has subtle effects on cardiovascular health, which can be observed through ECG analysis.
Article
Social Sciences, Interdisciplinary
Luis B. Elvas, Joao C. Ferreira, Miguel Sales Dias, Luis Bras Rosario
Summary: This paper proposes a novel approach to health data access in a hospital environment, leveraging homomorphic encryption to ensure privacy and secure sharing of medical data among healthcare entities. The framework establishes a secure environment that enforces GDPR adoption and integrates artificial intelligence capabilities for data analysis. Machine learning models achieve high accuracies of 96.88% with logistic regression and 97.62% with random forest. The comprehensive framework combines homomorphic encryption, secure data sharing, and AI-driven analytics, contributing to the advancement of medical research and healthcare outcomes.
Article
Engineering, Electrical & Electronic
Mariana C. Jacob Rodrigues, Octavian Postolache, Francisco Cercas
Summary: This study examines the effects of sound stimulation, such as music and stress noise, on the balance of the autonomous nervous system. The results showed that ambient music increases parasympathetic activity and comfort levels, while noise stress contributes to the increase of sympathetic activity. Integrating musical stimuli into a smart environment can potentially lower individual's stress levels and improve well-being.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Information Systems
Bruno Mataloto, Joao C. Ferreira, Ricardo Pontes Resende
Summary: The Internet of Things (IoT) allows real-time monitoring of energy consumption in smart homes using embedded sensors. This study proposes an innovative approach using color-based dashboards to enhance user interaction and achieve long-term energy savings. The approach includes management of appliances and comfort levels based on user preferences. Multiple strategies such as 3D representation and mobile connectivity are implemented to increase user engagement. The approach achieved an average energy consumption reduction of 19% and sustained user engagement over time, with participants considering it more attractive and useful than existing solutions based on a community survey.
Article
Computer Science, Information Systems
Luis B. Elvas, Pedro Aguas, Joao C. Ferreira, Joao Pedro Oliveira, Miguel Sales Dias, Luis Bras Rosario
Summary: This study investigates the efficacy of five CNN models combined with transfer learning and data augmentation techniques in accurately classifying AS. The VGG16 model achieves high recall and F1-score, and various data augmentation techniques are implemented to improve the model's robustness. The validation results confirm the clinical applicability of the model in real cases.