Search in Publications
Found 6278 publications
Uncertainty Quantification in Shear Wave Velocity Predictions: Integrating Explainable Machine Learning and Bayesian Inference
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
NFC applications and business model of the ecosystem
Publication Name: 2007 16th Ist Mobile and Wireless Communications Summit
Publication Date: 2007-12-01
Volume: Unknown
Issue: Unknown
Page Range: Unknown
Description:
StoLPaN, a pan-European consortium of companies, universities and user groups co-funded by the European Commission (EU), Information Society Technologies (IST) Programme aims to define open commercial and technical frameworks for the remote management of NFC-enabled services on mobile devices. These frameworks will facilitate the deployment of NFC-enabled mobile applications across a wide range of vertical markets, regardless of the phone type and the nature of the services required. This paper introduces the business aspects of the NFC based service development and the technical infrastructure implementing the core of the NFC-enabled services.
Open Access: Yes
Sick-pay trends of the National Health Insurance Fund of Hungary between 1997 and 2017
Publication Name: Orvosi Hetilap
Publication Date: 2019-02-01
Volume: 160
Issue: Unknown
Page Range: 37-42
Description:
Introduction: The role of sick-pay is to compensate for loss of wage in case of incapacity for work, to ensure that there will be no break in the existential state of the incapacitated person. Aim: The purpose of our research was to examine data on sickness benefit and payroll data for the period 1997–2017. Data and methods: Our research was based on the data of the National Health Insurance Fund of Hungary, the Hungarian Central Statistical Office, the State Audit Office of Hungary and the Hungarian State Treasury as well as on the background reports of the European Commission's Social Protection Committee and the Organisation for Economic Co-operation and Development (OECD), and it is based on the legal environment of sick-pay. Results: In 1997, there were 119 000 of 3.558 million, in 1998 114 000 of 3.530 million, in 1999 115 000 of 3.433 million, in 2000 112 000 of 3.465 million, in 2006 100 000 of 3.523 million, in 2012 55 000 of 3.769 million entitled people on sick leave on average per day. In 2017, the number of entitled persons increased to 4.018 million, while the average number of sick days per day was 70 000. According to data from gender and age-based analysis, in most cases, women with childbearing were on sick-leave, the proportion of males was higher in relation to industrial accident. Between 2014 and 2016, the proportion of women on sick leave per day was 59–60%, while the major reason for sick-pay among men was the industrial accident. The distribution by age did not change significantly. In 2014–2015, the age-group 30–34, while in 2016 the age-group 35–39 had the highest utilization of sick-pay. Conclusion: We can conclude that the use of sick-pay is affected by the employment rate, legal changes affecting the amount of sick-pay and social trends like substitution difficulties due to labor shortages and fear of losing jobs.
Open Access: Yes
Optimization on physicomechanical and wear properties of wood waste filled poly(lactic acid) biocomposites using integrated entropy-simple additive weighting approach
Publication Name: South African Journal of Chemical Engineering
Publication Date: 2022-07-01
Volume: 41
Issue: Unknown
Page Range: 193-202
Description:
The present research work develops an evaluation method based on the hybrid entropy-simple additive weighting approach to select the best biocomposite material based on several potentially conflicting criteria. Poly(lactic acid) (PLA) biocomposites with varying proportions of wood waste (0, 2.5, 5, 7.5, and 10 by weight) was developed and evaluated for physical, mechanical, and sliding wear properties. The biocomposite containing 10 wt.% wood waste exhibited the lowest density (1.183 g/cm3) and highest modulus properties (tensile modulus = 2.97 GPa; compressive modulus = 3.46 GPa; and flexural modulus = 4.03 GPa). The bare PLA exhibited the highest strength properties (tensile strength = 57.96 MPa; compressive strength = 105.67 MPa; impact strength = 15.25 kJ/m2), whereas flexural strength (100.43 MPa) was the highest for 5 wt.% wood waste filled biocomposite. The wear of PLA decreased with 2.5 wt.% wood waste incorporated and increased with further addition of wood waste. The experimental results revealed a high compositional dependence with no discernible trend. As a result, prioritizing biocomposites' performance to choose the best from various biocomposite alternatives becomes tough. Therefore, a multi-criteria decision-making process based on a hybrid entropy-simple additive weighting approach was applied to find the optimal biocomposite by taking the experimental results as the selection criterion. The results show that the 2.5 wt.% wood waste added PLA biocomposite proved to be the best solution with optimal physical, mechanical, and wear properties. The validation with other decision-making models supports the robustness of the proposed approach in that the 2.5 wt.% wood waste added PLA biocomposite is the most dominating. This study contributes by providing preferences for the selection criteria and assessing the best alternative from the available PLA biocomposites.
Open Access: Yes
Systematic Review of Cashierless Stores (Just Walk Out Stores) Revolutionizing The Retail
Publication Name: Management and Marketing
Publication Date: 2023-12-01
Volume: 18
Issue: s1
Page Range: 427-448
Description:
The paper aims to examine the evolving retail sector in recent years, specifically how digitalisation and technological innovations have transformed it. All actors have had to adapt to remain competitive. Notably, a new innovation in the retail sector, namely the checkout-free or cashierless store, emerged in 2018. Systematic literature is relied upon to achieve the study's objectives. The significance of this study lies in the use of multiple IT tools such as AI, cameras, sensors, and self-organising shelves to replace human intervention in the retail sector. Globally, several startup companies have developed this new unmanned solution, and Amazon Go stands out as one of the most well-known among them. The primary objective of this pioneering concept is to enhance efficiency by saving time and reducing queues. The aim is to enable customers to enter and exit the store with minimal human contact as quickly as possible. This paper presents the recent trend of the cashierless concept, its evolution, and proliferation. A systematic literature review and data analysis from the Crunchbase Database were conducted. The findings demonstrate that this recent concept is altering both consumers' purchasing behaviours and companies' business models. This paper provides novel perspectives and insights into the wider literature on cashierless concepts and smart retail in the context of digital business.
Open Access: Yes
Transferability of safety inspection procedures for network-wide safety assessment of two-lane rural roads - an Italian-Hungarian experiment
Publication Name: Traffic Injury Prevention
Publication Date: 2026-01-01
Volume: 27
Issue: 4
Page Range: 446-454
Description:
Objectives: The new EU Directive on Road Infrastructure Safety Management requires Member States to classify the road network into at least three categories according to its safety level. This study examines the application and transferability of the procedures between EU countries. Methods: Our methodology consisted of two steps. First, we conducted a questionnaire survey among twenty Hungarian road safety inspectors, and second, we applied the Italian procedure to calculate the risk index and compare it with historical crash data. Two-lane rural roads were selected and divided into 200 m sections, excluding intersections. Road safety inspectors evaluated these using a matrix of 18 criteria based on video recordings. The risk index was calculated, together with a sensitivity analysis, and its consistency with the observed crash history was investigated. Finally, three homogeneous groups were identified using k-medoids cluster analysis. Results: The survey showed good acceptance of the process, but we also found differences in how inspectors rated certain criteria. Our analysis of inspectors’ ratings of severity showed that there were varying degrees of agreement. However, we also concluded that the three-level rating may help to reduce disagreement. Our risk index calculations used four years of crash data, and a moderate correlation between the crash rate and the risk index was found. By assigning a weighted average of adjacent sections and performing a k-medoids cluster analysis, we found that the optimal number of clusters is three, and these show a meaningful relationship with crash frequency. Conclusion: Regarding the application of the Italian procedure in Hungary to meet the requirements of the new EU RISM, the results are promising, and the lessons learned may also be useful for other countries.
Open Access: Yes
Lessons to be learned in adoption of autonomous equipment for field crops
Publication Name: Applied Economic Perspectives and Policy
Publication Date: 2022-06-01
Volume: 44
Issue: 2
Page Range: 848-864
Description:
Autonomous equipment for crop production is on the verge of technical and economic feasibility, but government regulation may slow its adoption. Key regulatory issues include requirements for on-site human supervision, liability for autonomous machine error, and intellectual property in robotic learning. As an example of the impact of regulation on the economic benefits of autonomous crop equipment, analysis from the United Kingdom suggests that requiring 100% on-site human supervision almost wipes out the economic benefits of autonomous crop equipment for small and medium farms and increases the economies-of-scale advantage of larger farms.
Open Access: Yes
DOI: 10.1002/aepp.13177
The Race to Sustainability: Decoding Green University Rankings Through a Comparative Analysis (2018–2022)
Publication Name: Innovative Higher Education
Publication Date: 2025-02-01
Volume: 50
Issue: 1
Page Range: 241-275
Description:
This study investigates the evolving landscape of green universities by analyzing and comparing rankings from 2018 to 2022. It expands beyond the single score offered by the UI GreenMetric, employing Multi-Criteria Decision-Making (MCDM) techniques to evaluate universities from diverse perspectives. Focusing on the top 50 universities from 2022, the study assesses their performance across six key criteria: setting and infrastructure, energy and climate change, waste, water, transportation, and education and research. Various MCDM methods (LOPCOW MEREC, CoCoSo, CRADIS, EDAS, MABAC, MAIRCA, and MARCOS) are implemented, revealing how they prioritize different aspects of sustainability. Furthermore, the study examines the correlation between rankings and employs the COPELAND aggregation approach to derive a unified ranking. This investigation not only contrasts MCDM outcomes with the UI GreenMetric’s total score-based rankings but also illuminates the relative significance of each criterion and its variation across weighting techniques. Additionally, the study delves into the temporal dynamics of university rankings, offering insights into institutional performance across different years.
Open Access: Yes
Real-world performance analysis of a universal computational reasoning model for precision oncology in lung cancer
Dóra Mathiász
Christophe Le Tourneau
Dóra Tihanyi
Ákos Boldizsár
Júlia Déri
Maud Kamal
Dóra Lakatos
István Vályi-Nagy
László Urbán
Barbara Vodicska
Dóra Kormos
Anna Dirner
Róbert Dóczi
Réka Szalkai-Dénes
Ákos Takács
Mária Kocsis-Steinbach
Gábor György Kalmár
Márton Bolyácz
Viktor Kardos
Christian Rolfo
Gábor Pajkos
István Peták
Richárd Schwáb
Edit Várkondi
Arkadiusz Z. Dudek
Publication Name: Npj Precision Oncology
Publication Date: 2025-12-01
Volume: 9
Issue: 1
Page Range: Unknown
Description:
Tumors harbor multiple genetic alterations, yet treatment decisions are commonly based on single biomarkers, leading to underutilization of genomic information by comprehensive molecular tests, uncertainty in clinical practice, and frequent treatment failures. Although molecular tumor boards can assist personalized treatments, this process is not scalable or standardized, resulting in highly discordant recommendations. Validated digital solutions for personalized decision support are highly needed. The Digital Drug Assignment (DDA) system is a computational reasoning model that scores treatment options based on the full tumor genomic data. We retrospectively analyzed data of 111 lung cancer patients and found that high-score MTAs (1000≦DDA score) provided significant clinical benefit over other treatments, in terms of ORR, PFS, and OS. These results demonstrate that the DDA system is predictive of relative benefit of the various agents used in lung cancer care. Digital drug assignment can potentially address challenges with complex molecular profiles in routine clinical settings.
Open Access: Yes
