Article
Computer Science, Artificial Intelligence
Reza Ghezelbash, Abbas Maghsoudi, Mehdi Shamekhi, Biswajeet Pradhan, Mehrdad Daviran
Summary: In this study, unsupervised clustering and supervised machine learning methods were used to construct mineral prospectivity maps and genetic algorithm was incorporated to optimize model performance. The results showed that the genetic-based SVM model outperformed other models in detecting favorable areas associated with porphyry-type Cu mineralization.
NEURAL COMPUTING & APPLICATIONS
(2023)
Article
Acoustics
Shahin Sohrabi, Teresa Pamies Gomez, Jordi Romeu Garbi
Summary: The effectiveness of an active noise barrier relies on the placement of secondary sources and error sensors. This paper utilizes a genetic optimizer to find optimal locations for transducers based on specific criteria. Two approaches, the Two-step approach and the Multi-parameter approach, are employed to optimize the active noise control parameters. Results show that the Multi-parameter approach achieves better outcomes with less computational effort. The best configuration for the active noise barrier is determined to be control sources and error microphones placed at a height below the barrier's edge and distributed with an interval between a half and a full wavelength.
JOURNAL OF LOW FREQUENCY NOISE VIBRATION AND ACTIVE CONTROL
(2023)
Article
Meteorology & Atmospheric Sciences
P. L. Houtekamer, Bin He, Dominik Jacques, Ron McTaggart-Cowan, Leo Separovic, Paul A. Vaillancourt, Ayrton Zadra, Xingxiu Deng
Summary: The integration of short-range forecasts with a NWP model is an important step in an ensemble Kalman filter algorithm, with a multiphysics approach used in the Canadian global EnKF system. The study explores whether multiple integrations with different model physics versions can lead to more accurate and reliable probability distributions for model parameters. An evolutionary algorithm is employed to duplicate member configurations that contribute most to ensemble quality, while making slight perturbations to improve the system. The optimized system shows slight reductions in biases for humidity-sensitive radiance measurements and modest improvements in medium-range ensemble forecasts.
MONTHLY WEATHER REVIEW
(2021)
Article
Computer Science, Artificial Intelligence
Remy Estran, Antoine Souchaud, David Abitbol
Summary: This article demonstrates how an expert credit rating model can be optimized using a genetic algorithm, combining expert intelligence with artificial intelligence. The proposed model meets European banking regulatory requirements and achieves transparency and simplicity.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Biochemical Research Methods
Shima Shafiee, Abdolhossein Fathi, Ghazaleh Taherzadeh
Summary: Peptide-binding proteins play important roles in various applications. SPPPred is a novel ensemble machine learning-based approach that can predict protein-peptide binding residues with consistent and comparable performance.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Vinicius A. Paiva, Murillo Mendonca, Sabrina A. Silveira, David B. Ascher, Douglas E. Pires, Sandro C. Izidoro
Summary: This study presents a new method, Genetic Active Site Search (GASS)-Metal, for predicting metal-binding sites in proteins. The method uses a parallel genetic algorithm to find candidate metal-binding sites similar to curated templates, and has been thoroughly validated with satisfactory performance.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Biochemical Research Methods
Ran Liu, Ye-Fan Hu, Jian-Dong Huang, Xiaodan Fan
Summary: A Bayesian model is proposed to jointly infer core locations, binding affinity, and binding thresholds, providing accurate determination for each MHC. Simulation studies showed desirable estimation accuracy and robustness of the model, outperforming commonly used thresholds when applied to real data.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Engineering, Chemical
Daniel-Constantin Anghel, Daniela Monica Iordache, Alin Daniel Rizea, Nicolae-Doru Stanescu
Summary: This paper presents a study on obtaining certain clearance values for revolute joints of non-assembly mechanisms manufactured with FDM 3D Printing. By utilizing an artificial neural network and genetic algorithms, the relationship between imposed and measured clearances was explored to ensure assembly functionality. Experimental data was used for training the network, and a minimum clearance value was established through optimization, with a subsequent validation experiment confirming the results.
Article
Fisheries
Simon H. Fischer, Jose A. A. De Oliveira, John D. Mumford, Laurence T. Kell
Summary: Genetic algorithm can optimize catch rules and improve performance, with parameterization and improvement depending on specific stock and status. It is an efficient and automated method that can be applied to management procedures and case-specific tuning for fisheries.
ICES JOURNAL OF MARINE SCIENCE
(2021)
Article
Medicine, Research & Experimental
Cory A. Brennick, Mariam M. George, Marmar M. Moussa, Adam T. Hagymasi, Sahar Al Seesi, Tatiana V. Shcheglova, Ryan P. Englander, Grant L. J. Keller, Jeremy L. Balsbaugh, Brian M. Baker, Andrea Schietinger, Ion I. Mandoiu, Pramod K. Srivastava
Summary: Through an unbiased approach, a large number of effective anticancer neoepitopes have been identified, with properties distinct from conventional epitopes, offering potential for the development of personalized human cancer vaccines.
JOURNAL OF CLINICAL INVESTIGATION
(2021)
Article
Computer Science, Artificial Intelligence
Benjamin Alexander Albert, Yunxiao Yang, Xiaoshan M. M. Shao, Dipika Singh, Kellie N. N. Smith, Valsamo Anagnostou, Rachel Karchin
Summary: Researchers propose a method based on long short-term memory ensembles and transfer learning to predict effective neoepitopes that elicit an immune response. This method can help address the challenge of predicting immunogenicity of neoepitopes in developing personalized cancer vaccines. Compared to other state-of-the-art classifiers, this method significantly improves the prediction of epitope presentation.
NATURE MACHINE INTELLIGENCE
(2023)
Article
Forestry
Tu X. Ho, Laurence R. Schimleck, Arijit Sinha
Summary: The study developed an optimization problem using a genetic algorithm to select wavelengths for establishing multivariate calibration models based on partial least squares (PLS) regression, improving the performance of predicting Eucalyptus globulus pulp yield. The optimal number of latent variables ranged widely, from the maximum allowed to a lower limit, with representative wavelengths determined and assigned to corresponding wood components through a band assignment process.
WOOD SCIENCE AND TECHNOLOGY
(2021)
Article
Physics, Multidisciplinary
Feng Xu, Ling-Yu Mo, Hong Chen, Jia-Ming Zhu
Summary: For high-temperature clothing design, a heat transfer model was constructed using theories such as partial differential equations and finite difference method. The optimal fabric thickness parameters were obtained based on a human body burn model. The BP neural network weights and thresholds were optimized using a genetic algorithm to approximate and optimize the nonlinear thickness function with MATLAB.
FRONTIERS IN PHYSICS
(2021)
Article
Biochemical Research Methods
Zahed Khatooni, Navid Teymourian, Heather L. Wilson
Summary: This study introduces a novel strategy for epitope prediction using molecular dynamics simulation, homology modeling, and docking simulations. By selecting diverse SLA-1 alleles, the study successfully identifies virus epitopes that bind with high affinity to these alleles.
Article
Computer Science, Information Systems
Caroline A. Figueroa, Adrian Aguilera, Bibhas Chakraborty, Arghavan Modiri, Jai Aggarwal, Nina Deliu, Urmimala Sarkar, Joseph Jay Williams, Courtney R. Lyles
Summary: This study describes the challenges, considerations, and solutions for algorithm design decisions in a collaboration between health services researchers, clinicians, and data scientists. Nine challenges emerged, divided into three major themes: model selection for decision-making, data handling/collection, and weighing algorithm performance vs effectiveness/implementation in real-world settings.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2021)