Adaptive Neural Architecture Search Using Meta-Heuristics: Discovering Fine-Tuned Predictive Models for Photocatalytic CO2 Reduction DOI Open Access
Ümit Işıkdağ, Gebrai̇l Bekdaş, Yaren Aydın

и другие.

Sustainability, Год журнала: 2024, Номер 16(23), С. 10756 - 10756

Опубликована: Дек. 8, 2024

This study aims to contribute the reduction of carbon dioxide and production hydrogen through an investigation photocatalytic reaction process. Machine learning algorithms can be used predict yield in Although regression-based approaches provide good results, accuracy achieved with classification is not very high. In this context, presents a new method, Adaptive Neural Architecture Search (NAS) using metaheuristics, improve capacity ANNs estimating process classification. The NAS was carried out tool named HyperNetExplorer, which developed aim finding ANN architecture providing best prediction changing hyperparameters, such as number layers, neurons each layer, activation functions layer. nature adaptive, since accomplished optimization algorithms. discovered HyperNetExplorer demonstrated significantly higher performance than classical ML results indicated that helped achieve better estimation

Язык: Английский

Shear Wave Velocity Prediction with Hyperparameter Optimization DOI Creative Commons
Gebrai̇l Bekdaş, Yaren Aydın, Ümit Işıkdağ

и другие.

Information, Год журнала: 2025, Номер 16(1), С. 60 - 60

Опубликована: Янв. 16, 2025

Shear wave velocity (Vs) is an important soil parameter to be known for earthquake-resistant structural design and determining the dynamic properties of soils such as modulus elasticity shear modulus. Different Vs measurement methods are available. However, these methods, which costly labor intensive, have led search new Vs. This study aims predict (Vs (m/s)) using depth (m), cone resistance (qc) (MPa), sleeve friction (fs) (kPa), pore water pressure (u2) N, unit weight (kN/m3). Since varies with depth, regression studies were performed at depths up 30 m in this study. The dataset used open-source dataset, data from Taipei Basin. was extracted, a 494-line created. In study, HyperNetExplorer 2024V1, prediction based on shell (fs), (kN/m3) values could satisfactory results (R2 = 0.78, MSE 596.43). Satisfactory obtained Explainable Artificial Intelligence (XAI) models also used.

Язык: Английский

Процитировано

1

Predicting mechanical properties of CFRP composites using data-driven models with comparative analysis DOI Creative Commons
Ammar Alsheghri, A. H. Al-Hammadi,

Vassilis Drakonakis

и другие.

PLoS ONE, Год журнала: 2025, Номер 20(4), С. e0319787 - e0319787

Опубликована: Апрель 7, 2025

Carbon fiber reinforced polymer (CFRP) composites are increasingly utilized for their lightweight and superior mechanical properties. This study uses machine learning models to predict the properties of CFRP based on volume fraction carbon nanotubes (CNTs), interlayer fraction, glass transition temperature, manufacturing pressure. Sixty-two samples covering nine different types CFRPs were designed, manufactured, experimentally tested. Three models, namely ridge regression, random forest, support vector trained data compared. The results demonstrated a high prediction accuracy flexural strength (R 2 = 0.966), modulus 0.871), mode-II energy release rate 0.903). highlights effectiveness data-driven in predicting key composites, potentially reducing need extensive experimental testing facilitating more efficient material design.

Язык: Английский

Процитировано

1

Fatigue Predictive Modeling of Composite Materials for Wind Turbine Blades Using Explainable Gradient Boosting Models DOI Open Access
Yaren Aydın, Celal Çakıroğlu, Gebrai̇l Bekdaş

и другие.

Coatings, Год журнала: 2025, Номер 15(3), С. 325 - 325

Опубликована: Март 11, 2025

Wind turbine blades are subjected to cyclic loading conditions throughout their operational lifetime, making fatigue a critical factor in design. Accurate prediction of the performance wind is important for optimizing design and extending lifespan energy systems. This study aims develop predictive models laminated composite life based on experimental results published by Montana State University, Bozeman, Composite Material Technologies Research Group. The have been trained dataset consisting 855 data points. Each point consists stacking sequence, fiber volume fraction, stress amplitude, frequency, laminate thickness, number cycles test carried out specimen. output feature cycles, which indicates Random forest (RF), extreme gradient boosting (XGBoost), categorical (CatBoost), light machine (LightGBM), extra trees regressor predict specimens. For optimum performance, hyperparameters these were optimized using GridSearchCV optimization. total failure could be predicted with coefficient determination greater than 0.9. A importance analysis was SHapley Additive exPlanations (SHAP) approach. LightGBM showed highest among (R2 = 0.9054, RMSE 1.3668, MSE 1.8682).

Язык: Английский

Процитировано

0

PREDICTION OF THE COMPRESSIVE AND TENSILE STRENGTH OF HIGH-PERFORMANCE CONCRETE BASED ON A HYBRID MODEL OF MULTILAYER PERCEPTRON (MLP) AND LIGHTGBM DOI Creative Commons
S. J. Zhao

Ceramics - Silikaty, Год журнала: 2025, Номер unknown, С. 0 - 0

Опубликована: Май 13, 2025

Язык: Английский

Процитировано

0

Predicting compressive and tensile strength of concrete with different sand types using machine learning DOI

Tarek Salem Abdennaji,

Rupesh Kumar Tipu, Yahya Alassaf

и другие.

Ain Shams Engineering Journal, Год журнала: 2025, Номер 16(8), С. 103474 - 103474

Опубликована: Май 14, 2025

Язык: Английский

Процитировано

0

Predicting the area moment of inertia of beam and column using machine learning and HyperNetExplorer DOI
Yaren Aydın, Sinan Melih Niğdeli, Mostafa Roozbahan

и другие.

Neural Computing and Applications, Год журнала: 2025, Номер unknown

Опубликована: Май 30, 2025

Язык: Английский

Процитировано

0

Prediction of autogenous shrinkage in ultra-high-performance concrete (UHPC) using hybridized machine learning DOI
Md Muzammal Hoque,

Ajad Shrestha,

Sanjog Chhetri Sapkota

и другие.

Asian Journal of Civil Engineering, Год журнала: 2024, Номер unknown

Опубликована: Окт. 30, 2024

Язык: Английский

Процитировано

2

Adaptive Neural Architecture Search Using Meta-Heuristics: Discovering Fine-Tuned Predictive Models for Photocatalytic CO2 Reduction DOI Open Access
Ümit Işıkdağ, Gebrai̇l Bekdaş, Yaren Aydın

и другие.

Sustainability, Год журнала: 2024, Номер 16(23), С. 10756 - 10756

Опубликована: Дек. 8, 2024

This study aims to contribute the reduction of carbon dioxide and production hydrogen through an investigation photocatalytic reaction process. Machine learning algorithms can be used predict yield in Although regression-based approaches provide good results, accuracy achieved with classification is not very high. In this context, presents a new method, Adaptive Neural Architecture Search (NAS) using metaheuristics, improve capacity ANNs estimating process classification. The NAS was carried out tool named HyperNetExplorer, which developed aim finding ANN architecture providing best prediction changing hyperparameters, such as number layers, neurons each layer, activation functions layer. nature adaptive, since accomplished optimization algorithms. discovered HyperNetExplorer demonstrated significantly higher performance than classical ML results indicated that helped achieve better estimation

Язык: Английский

Процитировано

1