Search Everything

Tip: Search using "First Name + Last Name", e.g.
János Kiss instead of Kiss János.

Publications - 6674

Uncertainty Quantification in Shear Wave Velocity Predictions: Integrating Explainable Machine Learning and Bayesian Inference

Ayele Chala Richard P. Ray

Publication Name: Applied Sciences Switzerland

Publication Date: 2025-02-01

Volume: 15

Issue: 3

Page Range: Unknown

Description:

The accurate prediction of shear wave velocity (Vs) is critical for earthquake engineering applications. However, the prediction is inevitably influenced by geotechnical variability and various sources of uncertainty. This paper investigates the effectiveness of integrating explainable machine learning (ML) model and Bayesian generalized linear model (GLM) to enhance both predictive accuracy and uncertainty quantification in Vs prediction. The study utilizes an Extreme Gradient Boosting (XGBoost) algorithm coupled with Shapley Additive Explanations (SHAPs) and partial dependency analysis to identify key geotechnical parameters influencing Vs predictions. Additionally, a Bayesian GLM is developed to explicitly account for uncertainties arising from geotechnical variability. The effectiveness and predictive performance of the proposed models were validated through comparison with real case scenarios. The results highlight the unique advantages of each model. The XGBoost model demonstrates good predictive performance, achieving high coefficient of determination ((Formula presented.)), index of agreement (IA), Kling–Gupta efficiency (KGE) values, and low error values while effectively explaining the impact of input parameters on Vs. In contrast, the Bayesian GLM provides probabilistic predictions with 95% credible intervals, capturing the uncertainty associated with the predictions. The integration of these two approaches creates a comprehensive framework that combines the strengths of high-accuracy ML predictions with the uncertainty quantification of Bayesian inference. This hybrid methodology offers a powerful and interpretable tool for Vs prediction, providing engineers with the confidence to make informed decisions.

Open Access: Yes

DOI: 10.3390/app15031409

Develping artificial intelligence technology to support cattle identification, animal health and welfare solutions

István Szabó Tőzsér János Tarr Bence

Publication Name: Magyar Allatorvosok Lapja

Publication Date: 2023-01-01

Volume: 145

Issue: 11

Page Range: 651-660

Description:

Artificial Intelligence (AI) has become an important tool for optimising breeding processes in several areas of animal production. In this thesis, we have presented examples from the literature, mainly for the identification and counting of cattle. The individual identification of animals, the monitoring of their behaviour and the control of their movements support a number of conclusions from both animal welfare and veterinary point of view. Automation of the processing of captured images has also become essential. This process is supported by Artificial Intelligence. Deep learning and neural networks are excellent tools for segmenting images and processing their content based on different features. Convolutional neural networks are specifically powerful for such tasks and we have seen that further developments of these networks (e.g. Faster R-CNN) allow even more efficient image analysis procedures. Processing animal images can be a major step forward for automatic analysis and identification of livestock. It also allows early intervention in the event of disease. In the context of individual identification, it is important to underline that, when complemented with other measurement options, e.g. sensor measurements, it offers even more complex applications that have not been available so far.

Open Access: Yes

DOI: 10.56385/magyallorv.2023.11.651-660

Possibilities of Using of Online Vehicle Diagnostics in the Future

Istvan Lakatos Jozsef Nagy

Publication Name: Lecture Notes in Mechanical Engineering

Publication Date: 2023-01-01

Volume: Unknown

Issue: Unknown

Page Range: 71-83

Description:

In the premium vehicle category, real-time online internet connection has become a standard in recent years. This trend is likely to spread completely in the automotive industry in the coming years. This fact offers a lot of new options in the field of vehicle maintenance (and predictive maintenance). Another possible use case may be remote diagnostics of in-use vehicles on the market, analysis of their online data and thereby an extension of the product development process after SOP. An additional new option may be to automatically collect, evaluate and generate of onboard diagnostics data to report to different authorities. E.g. OBFCM (onboard fuel consumption) or IUMPR j3 (in use monitoring performance ratio) field reports. In vehicle production, during the test drive, it could be possible to read and log of measurement data of the finished vehicle’s control units online Another application may be to test vehicles online during the production process e.g. to read of DTC’s (diagnostics trouble codes) during technical tests or to monitor of SoC (state of charge) of battery online while moving vehicles within the factory.

Open Access: Yes

DOI: 10.1007/978-3-031-15211-5_7

Implications of climate-change-induced stressors and water management for sterlet populations in the Middle and Upper River Danube

Bořek Drozd Ladislav Pekárik Dušan Senko Maroš Kubala Martin Mišík Martin Šindler Gábor Guti

Publication Name: River Research and Applications

Publication Date: 2025-02-01

Volume: 41

Issue: 2

Page Range: 448-465

Description:

Sturgeons are a group of iconic rheophilic fish whose populations worldwide are currently undergoing significant declines. The study investigates the impact of climate change and anthropogenic activities, particularly the Gabčíkovo barrage system, on the survival and distribution of the sterlet, the last surviving sturgeon species in the Middle and Upper Danube River, specifically in the river part rkm 1708–1920 divided into three river sections: PT1 (Danube river part 1: rkm 1708–Gabčíkovo impoundment), PT2 (Danube river part 2: rkm 1854–1920), and PT3 (Old Danube part 3: rkm 1850–1810). Between years 1996 and 2020, water temperatures in the Danube River (PT1, PT2) rose by over 1°C on average, with the Old Danube (PT3) experiencing an alarming average increase of 3.5°C (df = 2, F = 145.03, p = <2.2e-16). Consequently, suitable sterlet habitat (depth ≥4 m) in the Old Danube (PT3) now covers only 11% of its total area compared with 67% in PT1 and 75% in PT2 (flow rate = 1924.27 m3 s−1) due to altered flow regimes, water levels, and siltation. Sterlets are increasingly caught in an artificial channel below to the Gabčíkovo Hydroelectric Power Plant (80%–90% of annual total catch), suggesting a shift in their distribution patterns. Conversely, there has been a notable decline in sterlet populations in other river sections, including a decrease in a section with moderate ecological status according to Water Frame Directive criteria. In light of these findings, the study proposes several mitigation measures to improve the status of the sterlet population in the area.

Open Access: Yes

DOI: 10.1002/rra.4361

How do economies decarbonize growth under finance-energy inequality? Global evidence

Gagan Deep Sharma Aviral Kumar Tiwari Hai Hong Trinh Diem Thi Hong Vo

Publication Name: Energy Economics

Publication Date: 2025-02-01

Volume: 142

Issue: Unknown

Page Range: Unknown

Description:

The study investigates the multidecade complexity between economic growth and carbon emissions across income groups and regions for 180 economies over the past decades. We find that the global economy has been decarbonizing its economic growth. The effects of growth on decarbonization are conditional on outcome distributions. The Paris Agreement (COP21) and renewable energy consumption (REC) are robust mechanisms toward green growth. Financial development (FD) presents its moderation to decarbonized growth. The study makes the following novel contributions to prior literature streams. First, complex GDP-CO2 nexuses are conditional on green factors and decarbonization is foremost for our global inclusive growth. Second, the friendliness of FD to the environment relies on green transition. It is worth noting that financial institutions and markets are exposed to climate risk drivers leading to our great challenge to promote green finance. Decarbonization is our global and constant efforts toward inclusive growth. Under finance-energy inequality, renewable energy capacity and finance are critical to decarbonized economic growth.

Open Access: Yes

DOI: 10.1016/j.eneco.2024.108172

Analysis of the Effect of Mixed Eccentricity Fault on Controlled Induction Machines via Finite Element Method

Gyorgy Istenes Krisztian Horvath

Publication Name: 2025 19th International Conference on Electrical Machines Drives and Power Systems Elma 2025 Proceedings

Publication Date: 2025-01-01

Volume: Unknown

Issue: Unknown

Page Range: Unknown

Description:

The eccentricity in induction machines is a geometrical fault when the axis of rotation deviates from the ideal. As a result, the air gap between the stator and the rotor varies. Since the eccentricity is caused by a geometric misalignment, finite element method is used to model the fault in this paper and the phenomenon is detected by spectral analysis of the stator currents. In addition, the effect of eccentricity on the controlled drive is also investigated by numerical simulations.

Open Access: Yes

DOI: 10.1109/ELMA65795.2025.11083515

Structural Integrity Analysis of 3D-Printed PLA- and ABS-Reinforced Concrete Underwater Curing †

Szabolcs Fischer Szabolcs Szalai A. L. Nagy Hanna Csótár Gusztáv Baranyai Attila Nagy

Publication Name: Engineering Proceedings

Publication Date: 2025-01-01

Volume: 113

Issue: 1

Page Range: Unknown

Description:

The optimization and evaluation of 3D-printed polylactic acid (PLA) and ABS materials present a promising approach for enhancing the reinforcement of concrete elements, thereby advancing sustainable construction technologies. This study examines the degradation of the structural integrity of 3D-printed PLA- and ABS-reinforced concrete after 28 days of underwater curing. The research focuses on macroscopic analysis and microscopic characterization using a digital microscope and a high-resolution camera to investigate the crystalline structures formed during curing. The findings will offer valuable insights into the structural transformations occurring within concrete elements and potential interactions between concrete and PLA structures, paving the way for future civil engineering applications.

Open Access: Yes

DOI: 10.3390/engproc2025113045

Effect of Printing Parameters on the Tensile Mechanical Properties of 3D-Printed Thermoplastic Polyurethane †

László Lendvai Sándor K. Jakab

Publication Name: Engineering Proceedings

Publication Date: 2025-01-01

Volume: 113

Issue: 1

Page Range: Unknown

Description:

Thermoplastic polyurethane (TPU) filament was used to fabricate specimens through material extrusion (MEX)-based 3D printing technique with varying printing parameters. Nozzle diameters of 0.4 mm and 0.8 mm were used, while the printing infill orientation (also denoted as raster angle) was either parallel (0°) to the length of the specimens, perpendicular to it (90°), or at a 45° angle with alternating direction in each layer (±45°). Tensile tests were conducted to determine tensile strength, Young’s modulus, and elongation at break of the samples. The highest tensile strength was achieved using a 0.8 mm nozzle diameter and 0° raster angle, reaching 32.5 MPa, with a corresponding Young’s modulus of 145.8 MPa. Meanwhile, the sample with the lowest modulus (100.4 MPa) and tensile strength (17.8 MPa) was the one 3D-printed with a 0.4 mm nozzle and 90° raster angle.

Open Access: Yes

DOI: 10.3390/engproc2025113019

Sentiment Analysis of Marketplace Lending Platforms: A Study Based on Natural Language Processing

Jewel Kumar Roy

Publication Name: International Journal of Business Analytics

Publication Date: 2025-01-01

Volume: 12

Issue: 1

Page Range: Unknown

Description:

This study explores the link between user sentiment and credit risk on FinTech lending platforms using sentiment analysis techniques like Latent Dirichlet Allocation (LDA) and the Liu Hu method. Analyzing data from 2020 to 2023, findings reveal Kiva leads with 91.16% positive feedback and a 4.7-star rating but fewer reviews (617). LendingClub, with 1.58K reviews, has mixed sentiment (56.08% positive, 39.99% negative) and a lower rating (3.3 stars). Plenti achieves 58.33% positive sentiment but lower coherence, while Mintos balances sentiment (66.69% positive) with the largest review base (100K+). Results show platforms with higher positive sentiment and topic coherence mitigate credit risk more effectively, underscoring the value of user feedback in optimizing marketplace lending. The study offers actionable insights for FinTech stakeholders to improve app performance and user-centric financial solutions through effective sentiment analysis.

Open Access: Yes

DOI: 10.4018/IJBAN.393942