1. Intelligent hybrid optimization of WEDM for Ti-6Al-4V using Taguchi deep learning and ANOVA for enhanced surface integrity and prediction accuracyPattanam Ramamoorthy Kannan, S. Balasubramani, S. Dinesh, M. Arul, 2025, original scientific article Keywords: wire electrical discharge machining, Ti-6Al-4V, surface integrity, hybrid optimization, Taguchi with deep learning prediction Published in DiRROS: 23.06.2025; Views: 102; Downloads: 52
Full text (2,49 MB) This document has many files! More... |
2. |
3. On the potential of ensemble forecasting for the prediction of meteotsunamis in the Balearic Islands : sensitivity to atmospheric model parameterizationsBaptiste Mourre, A. Santana, A. Buils, L. Gautreau, Matjaž Ličer, Agusti Jansá, B. Casas, B. Amengual, Joaquín Tintoré, 2021, original scientific article Abstract: This study investigates the potential of ensemble forecasting using full realistic high-resolution nested atmosphere–ocean models for the prediction of meteotsunamis in Ciutadella (Menorca, Spain). The sensitivity of model results to the parameterizations of the atmospheric model is assessed considering the ten most significant recent meteotsunami events for which observations are available. Different schemes adapted to high-resolution Weather Research and Forecasting model simulations were used for the representation of cumulus, microphysics, planetary boundary layer and longwave and shortwave radiations. Results indicate a large spread of the ensemble simulations in terms of the final magnitude of the meteotsunamis. While the modeling system is shown to be able to realistically trigger tsunamigenic atmospheric disturbances in more than 90% of the situations, the small-scale characteristics of these disturbances are significantly modified with the change of parameterizations, leading to significant differences in the magnitude of the simulated sea-level response. No preferred set of parameterizations can be identified that leads to either the largest or the most realistic magnitudes in the majority of situations. Instead, the performance of a given set of parameterizations is found to change with the meteotsunami event under consideration. Importantly, the observed magnitude of the extreme sea-level oscillations lies within the range of a nine-member ensemble in 70% of the cases. This ensemble approach would then allow to generate a realistic range of possibilities in the majority of events, thus improving the realism of meteotsunami predictions compared to single deterministic forecasts. Keywords: meteotsunamis prediction, atmosphere, ocean modeling, ensemble forecasting, atmospheric model parameterizations Published in DiRROS: 19.07.2024; Views: 564; Downloads: 381
Full text (4,16 MB) This document has many files! More... |
4. Breast cancer risk based on adapted IBIS prediction model in Slovenian women aged 40-49 years : coul it be better?Tjaša Oblak, Vesna Zadnik, Mateja Krajc, Katarina Lokar, Janez Žgajnar, 2020, original scientific article Keywords: breast surgery, IBIS, prediction models, risk factors Published in DiRROS: 12.07.2024; Views: 544; Downloads: 206
Full text (286,49 KB) |
5. Breast cancer risk prediction using Tyrer-Cuzick algorithm with an 18-SNPs polygenic risk score in a European population with below-average breast cancer incidenceTjaša Oblak, Petra Škerl, Benjamin J. Narang, Rok Blagus, Mateja Krajc, Srdjan Novaković, Janez Žgajnar, 2023, original scientific article Abstract: Goals: To determine whether an 18 single nucleotide polymorphisms (SNPs) polygenic risk score (PRS18) improves breast cancer (BC) risk prediction for women at above-average risk of BC, aged 40-49, in a Central European population with BC incidence below EU average. Methods: 502 women aged 40-49 years at the time of BC diagnosis completed a questionnaire on BC risk factors (as per Tyrer-Cuzick algorithm) with data known at age 40 and before BC diagnosis. Blood samples were collected for DNA isolation. 250 DNA samples from healthy women aged 50 served as a control cohort. 18 BC-associated SNPs were genotyped in both groups and PRS18 was calculated. The predictive power of PRS18 to detect BC was evaluated using a ROC curve. 10-year BC risk was calculated using the Tyrer-Cuzick algorithm adapted to the Slovenian incidence rate (S-IBIS): first based on questionnaire-based risk factors and, second, including PRS18. Results: The AUC for PRS18 was 0.613 (95 % CI 0.570-0.657). 83.3 % of women were classified at above-average risk for BC with S-IBIS without PRS18 and 80.7 % when PRS18 was included. Conclusion: BC risk prediction models and SNPs panels should not be automatically used in clinical practice in different populations without prior population-based validation. In our population the addition of an 18SNPs PRS to questionnaire-based risk factors in the Tyrer-Cuzick algorithm in general did not improve BC risk stratification, however, some improvements were observed at higher BC risk scores and could be valuable in distinguishing women at intermediate and high risk of BC. Keywords: early breast cancer, polygenic risk score, risk prediction Published in DiRROS: 21.03.2024; Views: 959; Downloads: 281
Full text (1,54 MB) |
6. Coastal flood risk assessment : an approach to accurately map flooding through national registry-reported eventsErik Kralj, Peter Kumer, Cécil J. W. Meulenberg, 2023, original scientific article Keywords: sea flood prediction, flooding maps, climate change resilience, natural disaster registry, coastal inundation, flood-prone areas Published in DiRROS: 08.12.2023; Views: 1193; Downloads: 559
Full text (7,60 MB) This document has many files! More... |
7. Sensitivity analysis of RF+clust for leave-one-problem-out performance predictionAna Nikolikj, Michal Pluhacek, Carola Doerr, Peter Korošec, Tome Eftimov, 2023, published scientific conference contribution Keywords: automated performance prediction, autoML, single-objective black-box optimization, zero-shot learning Published in DiRROS: 13.11.2023; Views: 1059; Downloads: 777
Full text (4,94 MB) This document has many files! More... |
8. MsGEN : measuring generalization of nutrient value prediction across different recipe datasetsGordana Ispirova, Tome Eftimov, Sašo Džeroski, Barbara Koroušić-Seljak, 2023, original scientific article Abstract: In this study, we estimate the generalization of the performance of previously proposed predictive models for nutrient value prediction across different recipe datasets. For this purpose, we introduce a quantitative indicator that determines the level of generalization of using the developed predictive model for new unseen data not presented in the training process. On a predefined corpus of recipe embeddings from six publicly available recipe datasets (i.e., projecting them in the same meta-feature vector space), we train predictive models on one of the six recipe datasets and test the models on the rest of the datasets. In parallel, we define and calculate generalizability indexes which are numbers that indicate how generalizable a predictive model is i.e., how well will a predictive model learned on one dataset perform on another one not involved in the training. The evaluation results prove the validity of these indexes – their relation with the accuracy of the predictions. Further, we define three sampling techniques for selecting representative data instances that will cover all parts from the feature space uniformly (involving data from all datasets) and further will improve the generalization of a predictive model. We train predictive models with these generalized datasets and test them on instances from the six recipe datasets that are not selected and included in the generalized datasets. The results from the evaluation of these predictive models show improvement compared to the results from the predictive models trained on one recipe dataset and tested on the others separately. Keywords: ML pipeline, predictive modeling, nutrient prediction, recipe datasets Published in DiRROS: 25.09.2023; Views: 1061; Downloads: 594
Full text (3,27 MB) This document has many files! More... |
9. |
10. Impact of leaching on chloride ingress profiles in concreteAlisa Machner, Marie Helene Bjørndal, Aljoša Šajna, Nikola Mikanovic, Klaartje De Weerdt, 2022, original scientific article Abstract: To investigate the effect of leaching on chloride ingress profiles in concrete and mortar, we exposed concrete and mortar specimens for 90 and 180 days to two different exposure solutions: 3% NaCl, and 3% NaCl with KOH added to limit leaching. The solutions were replaced weekly. After exposure, we determined total chloride profiles to investigate the chloride ingress, and portlandite profiles to assess the extent of leaching. The results showed that leaching during exposure greatly affects the chloride ingress profiles in mortar and concrete. We found that leaching leads to considerably higher maximum total chloride content and deeper chloride penetration into the concrete than in the specimens where leaching was limited. We recommend therefore that leaching should be taken into account in standard laboratory testing and that more mechanistic service life models should be used to take into account the impact of leaching. Keywords: chloride ingress, service life prediction, leaching, concrete, portlandite, open access Published in DiRROS: 04.05.2023; Views: 950; Downloads: 686
Full text (1,17 MB) This document has many files! More... |