Article
Computer Science, Theory & Methods
Jesus Medina, Ronald R. Yager
Summary: The paper introduces an extension of OWA operators allowing greater flexibility and the use of different types of OWA operators on the same dataset. Various interesting examples have been studied and a sufficient condition for ensuring the monotonicity of the new flexible operator has been presented.
FUZZY SETS AND SYSTEMS
(2021)
Article
Chemistry, Physical
Ryne C. Johnston, Kun Yao, Zachary Kaplan, Monica Chelliah, Karl Leswing, Sean Seekins, Shawn Watts, David Calkins, Jackson Chief Elk, Steven Jerome, Matthew P. Repasky, John C. Shelley
Summary: Epik version 7 is a software program that uses machine learning to predict the pKa values and protonation state distribution of complex, druglike molecules. It achieves high accuracy, with median absolute and root mean square errors of 0.42 and 0.72 pKa units, respectively, across seven test sets. Epik version 7 also provides protonation states and recovers 95% of the most populated states compared to previous versions. It is fast, taking only an average of 47 ms per ligand, making it suitable for evaluating protonation states and generating compound libraries for exploration of chemical space.
JOURNAL OF CHEMICAL THEORY AND COMPUTATION
(2023)
Article
Computer Science, Information Systems
Kazunori Iwata, Hiroki Yamamoto, Kazushi Mimura
Summary: Shape matching with local descriptors is a fundamental scheme in shape analysis. Our extended scheme considers the correspondence of neighboring sampled points, addressing computational feasibility issues with a branch-and-bound algorithm, and demonstrates a more suitable matching compared to traditional methods.
IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS
(2021)
Article
Energy & Fuels
Jialing Ma, Lifei Yin, Lixia Ling, Riguang Zhang, Guochun Yan, Jianli Wang, Weimin Lu, Yi Li, Baojun Wang
Summary: The study predicts the properties of adamantane derivatives and explores the influence of different substitutes and isomers on density, net heat value, specific impulse, thermal stability, and oxidation stability using density functional theory and molecular dynamics simulation. Descriptors such as the C/H ratio and the gap between HOMO and LUMO are used to determine the properties of the derivatives. The structures with superior properties are obtained for ethyladamantane, propyladamantane, and butyladamantane. The structure of pentyladamantane with excellent properties is predicted based on its resemblance to adamantane derivatives with preferable properties. The thermal and oxidation stability of certain derivatives meet the requirements of aircraft. Guidance on predicting, screening, and assessing the structure and properties of adamantane derivatives is provided in the study.
Article
Meteorology & Atmospheric Sciences
Hwan-Jin Song, Soonyoung Roh, Juho Lee, Giung Nam, Eunggu Yun, Jongmin Yoon, Park Sa Kim
Summary: Stochastic weight averaging (SWA) was used to improve the radiation emulator in a numerical weather prediction model over Korea. SWA showed significant improvements in forecast errors and stable long-term forecast features.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Chemistry, Physical
Ranga Rohit Seemakurthi, Griffin Canning, Zhenwei Wu, Jeffrey T. Miller, Abhaya K. Datye, Jeffrey Greeley
Summary: First principles periodic density functional theory calculations, combined with detailed microkinetic modeling and experimental characterization, were used to investigate the structure sensitivity and key selectivity descriptors for nonoxidative propane dehydrogenation (PDH) on intermetallic alloys. The study showed that Pd-terminated steps on 1:1 PdIn surfaces have significantly higher rates and selectivity towards propylene formation compared to pure Pd steps and terraces, suggesting a potential strategy for identifying selective nonoxidative PDH alloy catalysts.
Article
Chemistry, Physical
Robin Lawler, Yao-Hao Liu, Nessa Majaya, Omar Allam, Hyunchul Ju, Jin Young Kim, Seung Soon Jang
Summary: The study introduced a novel method combining DFT and ML to predict pK(a) values, finding that KRR with Pipeline 3 produced the best prediction results, with important features including the number of hydrogen atoms and the degree of oxidation in the molecule.
JOURNAL OF PHYSICAL CHEMISTRY A
(2021)
Article
Computer Science, Artificial Intelligence
Yiyao Zhou, Rui Chen, Yiqiang Zhao, Xiding Ai, Guoqing Zhou
Summary: This paper proposes a novel point cloud denoising algorithm based on non-local self-similarity characteristics, which utilizes adaptive curvature threshold and structure-aware descriptor to preserve structure and achieve high-quality reconstruction. The experiments show that the algorithm outperforms state-of-the-art methods in terms of reconstruction accuracy and structure preservation.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Xiao Teng, Xiang Zhang, Zhigang Luo
Summary: This paper proposes a multi-scale local cues and hierarchical attention-based LSTM model for stock price trend prediction. Experiments confirm the superior performance of the proposed model compared to existing models.
Article
Physics, Multidisciplinary
Ting Han, Jie Li, Liping Liu, Fengyu Li, Lin-Wang Wang
Summary: By comparing several feature types, we found that MTP feature has the strongest ability in describing physical systems, which is helpful for the development of MLFF.
NEW JOURNAL OF PHYSICS
(2023)
Article
Economics
Oscar Oelrich, Mattias Villani, Sebastian Ankargren
Summary: We propose a method called local prediction pools, which combines the predictive distributions of a group of experts based on variables believed to be related to their predictive accuracy. This two-step process involves estimating the conditional predictive accuracy of each expert using a set of covariates and then combining their predictive distributions based on this local predictive accuracy. The caliper method is introduced as a simple, fast, and interpretable approach to estimate the local predictive accuracy of each expert. The results show that local prediction pools outperform the widely used optimal linear pools in macroeconomic forecasting and predicting daily bike usage for a bike rental company.
JOURNAL OF FORECASTING
(2023)
Article
Operations Research & Management Science
Che Xu, Wenjun Chang, Weiyong Liu
Summary: This paper proposes a data-driven decision model based on local two-stage weighted ensemble learning to improve decision performance using historical decision data. The model utilizes historical assessments to learn optimal weights for base classifiers (BCs) and criterion weights, achieving interpretable decision-making.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Computer Science, Information Systems
Amuthan Arjunan, R. Kaviarasan
Summary: Vehicular Ad hoc Network (VANET) is considered as a potential network for safe interaction, but handling NLOS nodes during emergencies remains a major concern. WDHPBDS effectively predicts NLOS nodes using weighted distance and hyperbolic properties, with experimental results showing better performance compared to other NLOS detection schemes.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Hua Yang, Chenting Gong, Kaiji Huang, Kaiyou Song, Zhouping Yin
Summary: In this study, a novel multi-bit binary descriptor method for face recognition is proposed, which utilizes a learned mapping matrix and multi-bit coding rules to extract multi-bit local binary descriptors. Additionally, a robust weight learning method is introduced to integrate these descriptors into the final face representation. Experimental results demonstrate the superior performance of the proposed method compared to state-of-the-art face recognition methods.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Materials Science, Multidisciplinary
Wan-Jian Yin
Summary: Descriptors are crucial in catalysis and the commonly used ones require experimental measurements or DFT calculations, hindering the design of new catalysts. A DFT-free descriptor derived via the SR approach provides a new solution to accelerate catalyst design.
COMPUTATIONAL MATERIALS SCIENCE
(2021)