Article
Multidisciplinary Sciences
Sung-Cheol Kim, Adith S. Arun, Mehmet Eren Ahsen, Robert Vogel, Gustavo Stolovitzky
Summary: Binary classification is a central problem in machine learning, and there is a surprising relationship between the probability of sample belonging to one class and the Fermi-Dirac distribution. The AUC is related to the temperature of an equivalent physical system, and the optimal decision threshold is associated with the chemical potential. A closed-form expression for calculating the variance of AUC has been derived, and an ensemble learning algorithm called FiDEL has been introduced.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Computer Science, Artificial Intelligence
A. Sujan Reddy, S. Akashdeep, R. Harshvardhan, S. Sowmya Kamath
Summary: This paper proposes a stacking ensemble model for short-term electricity consumption prediction. The experimental results show that the ensemble model, which combines predictions from multiple base models, achieves higher accuracy while reducing training time and root mean square error compared to existing techniques.
ADVANCED ENGINEERING INFORMATICS
(2022)
Article
Biochemistry & Molecular Biology
Gabriel Bianchin de Oliveira, Helio Pedrini, Zanoni Dias
Summary: Protein secondary structure prediction is vital in biological processes, with computational methods becoming the primary approach. By utilizing ensembles of different sub-classifiers, the accuracy of predictions can be improved.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Pawel Ksieniewicz, Pawel Zyblewski, Robert Burduk
Summary: Ensembles of classifiers are known for their stability and accuracy, often outperforming single classifiers. This study proposes a fusion method in geometric space using decision boundaries of base classifiers, introducing a new function for measuring central tendency and removing the limit on the number of base classifiers. Experiments on multiple binary datasets show the effectiveness of this approach.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
L. V. Rajani Kumari, Y. Chalapathi Rao
Summary: In this paper, a Pattern adaptive wavelet-based hybrid approach is proposed for classification of arrhythmia beats. The goal is to categorize Electrocardiogram (ECG) beats into normal and abnormal beats using various machine learning classification methods. Two hybrid classifiers using ensemble learning techniques are proposed to improve the performance. The proposed approach outperforms individual classifiers with increased accuracies.
SIGNAL IMAGE AND VIDEO PROCESSING
(2023)
Article
Engineering, Electrical & Electronic
Lucas Bastos, Geam Pfeiff, Ramon Oliveira, Helder Oliveira, Maria Emilia Tostes, Sherali Zeadally, Eduardo Cerqueira, Denis Rosario
Summary: Non-Technical Losses (NTL) pose a challenge due to prevalent electricity theft in traditional and smart metering systems. Proposed methods fail to accurately detect this fraud, making it difficult for companies to identify its origin and evaluate its financial impact. Combining time-series classification algorithms that consider the time-dependent nature of energy consumption data is essential. To address this, we propose DETECT, a data-oriented predictor for NTL detection that can effectively analyze and classify frauds based on consumption patterns, achieving a detection rate of 93.45% and a false positive rate of 1.61%.
ELECTRIC POWER SYSTEMS RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Rajkamal Rajendran, Anitha Karthi
Summary: This article describes a study on using machine learning to predict heart diseases in the healthcare industry. The authors propose a new machine learning pipeline, which includes preprocessing and feature engineering methods, and demonstrate through experiments that it outperforms other methods. The experimental results show significant improvement in accurately predicting heart diseases using the proposed pipeline, surpassing the current state-of-the-art results.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Chemistry, Analytical
Jaeyong Kang, Zahid Ullah, Jeonghwan Gwak
Summary: The study proposed a method for brain tumor classification using deep features and machine learning classifiers, adopting the concept of transfer learning and pre-trained deep convolutional neural networks. Experimental results demonstrated that an ensemble of deep features can significantly improve performance, with support vector machine outperforming other classifiers on large datasets.
Article
Mathematics, Interdisciplinary Applications
Matheus Henrique Dal Molin Ribeiro, Ramon Gomes da Silva, Gabriel Trierweiler Ribeiro, Viviana Cocco Mariani, Leandro dos Santos Coelho
Summary: Efficient models for short-term load forecasting in electricity distribution and generation systems are crucial for companies' energetic planning. In this study, an ensemble learning model based on dual decomposition approach, machine learning models and hyperparameters optimization is proposed. The model successfully decomposes the time series and handles the non-linearities, and achieves accurate load forecasting results with reduced errors.
CHAOS SOLITONS & FRACTALS
(2023)
Article
Management
Sandra Benitez-Pena, Emilio Carrizosa, Vanesa Guerrero, M. Dolores Jimenez-Gamero, Belen Martin-Barragan, Cristina Molero-Rio, Pepa Ramirez-Cobo, Dolores Romero Morales, M. Remedios Sillero-Denamiel
Summary: The paper introduces a novel Mathematical Optimization model for balancing the accuracy of the ensemble and the number of base regressors used by penalizing regressors with poor individual performance. The approach is shown to be flexible in incorporating various desirable properties one may have on the ensemble, such as controlling the performance in critical groups of records or managing costs associated with the base regressors.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2021)
Article
Mathematics
Fath U. Min Ullah, Noman Khan, Tanveer Hussain, Mi Young Lee, Sung Wook Baik
Summary: This study compares the performance of traditional and sequential learning algorithms in electricity load forecasting and presents a new ECP framework that utilizes deep sequential learning models to achieve optimal model optimization.
Article
Engineering, Civil
Guanjun Liu, Zhengyang Tang, Hui Qin, Shuai Liu, Qin Shen, Yuhua Qu, Jianzhong Zhou
Summary: This paper proposes a novel multi-model ensemble method, deep learning multi-dimensional ensemble method, to address the problem of multi-step runoff prediction. The method combines snapshot ensemble and attention ensemble techniques, significantly improving the runoff prediction performance. A deep learning multi-dimensional ensemble model is also proposed by combining three different deep learning neural networks with the ensemble method. The model is applied in a real-world case study and outperforms other comparison models, highlighting the effectiveness of the deep learning multi-dimensional ensemble method in hydrological predictions.
JOURNAL OF HYDROLOGY
(2022)
Article
Chemistry, Multidisciplinary
Manuel Lopez-Martin, Antonio Sanchez-Esguevillas, Luis Hernandez-Callejo, Juan Ignacio Arribas, Belen Carro
Summary: This study compares various traditional machine learning and deep learning techniques, as well as new methods for dynamic model analysis and short-term load forecasting. It explores the impact of critical parameters in time series forecasting, including rolling window length, forecast length, and the number/nature of features used.
APPLIED SCIENCES-BASEL
(2021)
Article
Engineering, Civil
Maria Kaiser, Stephan Guennemann, Markus Disse
Summary: This study developed a novel methodology based on tree-based classifiers to assess flood susceptibility at a regional scale using spatially distributed and catchment-related factors. The methodology was evaluated in the region of Bavaria (Germany), and all three models performed well, with the CatBoost model achieving the highest performance. It was found to be crucial to consider sample density and coverage of the study area when modeling large territories. The study also proposed an overall susceptibility score for cities based on the generated susceptibility map.
JOURNAL OF HYDROLOGY
(2022)
Article
Environmental Sciences
Connor J. Anderson, Daniel Heins, Keith C. Pelletier, Joseph F. Knight
Summary: This study examines the efficacy of a voting-based ensemble classifier to identify invasive Phragmites australis from imagery acquired from uncrewed aircraft systems. Results show that the voting-based ensemble classifier performs well in accurately identifying the invasive plant species, particularly when using multispectral imagery with an accuracy of 91%. The study highlights the need for further research on accurately identifying Phragmites australis at low stem densities.
Article
Materials Science, Characterization & Testing
Hao Wu, Guoyan Zhao, Pinnaduwa H. S. W. Kulatilake, Weizhang Liang, Enjie Wang
Article
Construction & Building Technology
Hao Wu, Guoyan Zhao, Weizhang Liang, Enjie Wang, Shaowei Ma
ADVANCES IN CIVIL ENGINEERING
(2019)
Article
Engineering, Mechanical
Hao Wu, Guoyan Zhao, Weizhang Liang
INTERNATIONAL JOURNAL OF MECHANICAL SCIENCES
(2019)
Article
Computer Science, Interdisciplinary Applications
Hao Wu, Pinnaduwa H. S. W. Kulatilake, Guoyan Zhao, Weizhang Liang, Enjie Wang
COMPUTERS AND GEOTECHNICS
(2019)
Article
Engineering, Mechanical
Hao Wu, Pinnaduwa H. S. W. Kulatilake, Guoyan Zhao, Weizhang Liang
THEORETICAL AND APPLIED FRACTURE MECHANICS
(2019)
Article
Engineering, Mechanical
Hao Wu, Guoyan Zhao, Weizhang Liang
THEORETICAL AND APPLIED FRACTURE MECHANICS
(2020)
Review
Geochemistry & Geophysics
Weizhang Liang, Bing Dai, Guoyan Zhao, Hao Wu
Article
Chemistry, Multidisciplinary
Hao Wu, Bing Dai, Guoyan Zhao, Ying Chen, Yakun Tian
APPLIED SCIENCES-BASEL
(2020)
Article
Mathematics
Weizhang Liang, Suizhi Luo, Guoyan Zhao, Hao Wu
Article
Geosciences, Multidisciplinary
Weizhang Liang, Asli Sari, Guoyan Zhao, Stephen D. McKinnon, Hao Wu
Article
Metallurgy & Metallurgical Engineering
Hao Wu, Bing Dai, Li Cheng, Rong Lu, Guoyan Zhao, Weizhang Liang
Summary: This study investigated the mechanical properties and energy evolution of sandstone samples with and without circular cavities under impact loading tests. The presence of a cavity weakened the dynamic compressive strength and affected crack initiation and rock fragmentation.
MINING METALLURGY & EXPLORATION
(2021)
Article
Geochemistry & Geophysics
Qingliang Chang, Xingjie Yao, Chongliang Yuan, Qiang Leng, Hao Wu
Summary: When the coal seam floor contains a confined aquifer, water inrush disasters are highly likely to occur. The failure behavior of the coal seam floor in the paste filling working face can be assessed through a theoretical model and numerical analysis, with the filling interval and long-term strength of the filling body having significant impacts on the floor failure depth, stress and displacement distributions, and plastic zone.
Article
Engineering, Civil
Hao Wu, Dan Ma, A. J. S. Spearing, Guoyan Zhao
Summary: Understanding the failure phenomena and mechanisms of stress regime on tunnels is critical, as hazardous failure phenomena frequently occur in deep hardrock tunnels. This study systematically investigated the mechanical behavior of rock specimens with different numbers of openings under uniaxial compression, revealing the significant impact of the number of openings on the weakening effect of rock mechanical properties. The crack evolution process in specimens with openings includes tensile cracks, spalling cracks, and shear cracks, and the crack initiation mechanism can be fully explained based on the stress distribution law around the opening.
GEOMECHANICS AND ENGINEERING
(2021)
Article
Computer Science, Information Systems
Shaowei Ma, Zhouquan Luo, Hao Wu, Yaguang Qin