Aksaray Üniversitesi Kurumsal Akademik Arşivi
DSpace@Aksaray, Aksaray Üniversitesi tarafından doğrudan ve dolaylı olarak yayınlanan; kitap, makale, tez, bildiri, rapor, araştırma verisi gibi tüm akademik kaynakları uluslararası standartlarda dijital ortamda depolar, Üniversitenin akademik performansını izlemeye aracılık eder, kaynakları uzun süreli saklar ve telif haklarına uygun olarak Açık Erişime sunar.
Güncel Gönderiler
Evaluation of endoscopic findings in gastrointestinal tract wall thickening detected on abdominal radiological imaging: a two-center retrospective descriptive study
(Multidisciplinary Digital Publishing Institute (MDPI), 2025) Ergin, Mustafa; Kıvrakoğlu, Fatih
Background and Objectives: The clinical significance of gastrointestinal (GI) tract wall thickening incidentally detected on abdominal imaging remains unclear. This study aimed to examine the relationship between GI tract wall thickening seen in imaging and what is found during endoscopy, as well as to explore how hemoglobin, C-reactive protein (CRP), and albumin levels can help predict the presence of malignancy. Materials and Methods: In this retrospectively designed study, 209 patients were included who were found to have GI tract wall thickening on radiological imaging and underwent endoscopy within 90 days. Endoscopic findings and laboratory data were recorded. Patients were compared based on the presence or absence of malignancy, and a receiver operating characteristic analysis was performed. Results: Malignancy was detected in 8.5% and 10.9% of the upper and lower GI tract cases, respectively. In patients with upper GI tract malignancy, hemoglobin levels were significantly lower and CRP levels were higher (p < 0.001 and p = 0.015, respectively). Similarly, in lower GI tract malignancy, hemoglobin levels were lower (p = 0.033), whereas CRP did not show a significant difference (p = 0.115). Cut-off values were determined as 11.8 g/dL for hemoglobin and 40.75 g/L for albumin, and both were found to have high negative predictive values. Conclusions: GI tract wall thickening detected radiologically is clinically significant and should be further investigated endoscopically. Certain biochemical parameters may aid in ruling out malignancy; however, endoscopy remains essential for definitive diagnosis.
Enhancing photovoltaic energy output predictions using ANN and DNN: a hyperparameter optimization approach
(Multidisciplinary Digital Publishing Institute (MDPI), 2025) Coşgun, Atıl Emre
This study investigates the use of artificial neural networks (ANNs) and deep neural networks (DNNs) for estimating photovoltaic (PV) energy output, with a particular focus on hyperparameter tuning. Supervised regression for photovoltaic (PV) direct current power prediction was conducted using only sensor-based inputs (PanelTemp, Irradiance, AmbientTemp, Humidity), together with physically motivated-derived features (ΔT, IrradianceEff, IrradianceSq, Irradiance × ΔT). Samples acquired under very low irradiance (<50 W m−2) were excluded. Predictors were standardized with training-set statistics (z-score), and the target variable was modeled in log space to stabilize variance. A shallow artificial neural network (ANN; single hidden layer, widths {4–32}) was compared with deeper multilayer perceptrons (DNN; stacks {16 8}, {32 16}, {64 32}, {128 64}, {128 64 32}). Hyperparameters were selected with a grid search using validation mean squared error in log space with early stopping; Bayesian optimization was additionally applied to the ANN. Final models were retrained and evaluated on a held-out test set after inverse transformation to watts. Test performance was obtained as MSE, RMSE, MAE, R2, and MAPE for the ANN and DNN. Hence, superiority in absolute/squared error and explained variance was exhibited by the ANN, whereas lower relative error was achieved by the DNN with a marginal MAE advantage. Ablation studies showed that moderate depth can be beneficial (e.g., two-layer variants), and a simple bootstrap ensemble improved robustness. In summary, the ANN demonstrated superior performance in terms of absolute-error accuracy, whereas the DNN exhibited better consistency with relative-error accuracy.
Effects of implementing interactive videos in an online flipped classroom on preservice special education teachers
(John Wiley and Sons Inc, 2025) Arslantaş, Tuğba Kamalı; Başer, Derya
This study investigates the impact of an online flipped classroom model, enriched with interactive videos, on preservice special education teachers' self-efficacy in using assistive technologies (AT) for visual impairment, their adoption of video-based learning, and their perceptions of interactive videos as effective learning tools. Addressing challenges such as limited access to AT tools and faculty expertise, the study explores how video-based instructional approaches can enhance teacher preparation for inclusive education. A mixed-methods intervention was conducted with 64 preservice teachers enrolled in an AT course at a university in Türkiye. Quantitative data were collected through pre- and post- intervention self-efficacy scales and a post-intervention adoption scale, while qualitative insights were gathered from weekly reflection reports and a final open-ended questionnaire. Results revealed a significant increase in preservice teachers' self-efficacies following the intervention. In addition, participants reported high levels of adoption of video-based learning, particularly in terms of perceived usefulness, enjoyment, ease of use, and intention to continue using interactive videos, suggesting a positive perception toward the integration of such tools in educational contexts. Qualitative findings reinforced the effectiveness of integrating interactive videos within an online flipped classroom to build AT-related competencies. The study offers practical implications for teacher education and suggests avenues for future research.
Prediction of cutting parameters and reduction of output parameters using machine learning in milling of Inconel 718 alloy
(Nature Research, 2025) Hamıd, Maher Waleed Hamıd; Özlü, Barış; Ulaş, Hasan Basri; Demir, Halil
The present study focuses on the effects of cooling/lubrication conditions and cutting parameters on energy consumption (EC), carbon emissions (CE), surface roughness (Ra), cutting temperature (T), tool wear (Vb) and vibration (Vib) in sustainable milling of Inconel 718 alloy. Also, it was aimed to estimate the EC, CE, Ra, T and Vib values obtained in milling experiments using three different regression-based machine learning (ML) models. The performances of the models used in ML were compared using R-squared, MSE and MAPE performance criteria. In the experiments conducted by reducing the feed rate and cutting speed and in MQL machining conditions, it was observed that EC and CE values reached minimum values. In MQL machining conditions, it was observed that the lowest Ra values were achieved at high cutting speed and low feed rate. The lowest Vb was measured at low cutting speed and feed rate in air machining conditions. Increasing the cutting speed and decreasing the feed rate in MQL machining conditions had a positive effect on Vib. In the MQL machining condition, at 40 m/min cutting speed and 0.06 mm/rev feed rate, the lowest energy consumption and carbon emission were 0.76 kJ/s and 0.54796 kg-CO2 respectively. The lowest surface roughness and vibration values were measured as 0.234 μm and 1.91 mm/s respectively, at 80 m/min cutting speed and 0.06 mm/rev feed rate in MQL machining condition. The lowest cutting temperature was measured as 31 °C at a cutting speed of 40 m/min and a feed rate of 0.06 mm/rev under air machining conditions. It was seen that the EC, CE, Ra, T and Vib values arising from the input parameters in the machining of Inconel 718 alloy could be successfully predicted using three different regression-based ML models.
Bringing exosomes into the game: Current situation, opportunities, limitations and future perspectives
(Elsevier Ltd, 2025) Dikici, Emrah; Önal Acet, Burcu; Gül, Désirée; Kummer, Nina; Stauber, Roland H.; Odabaşı, Mehmet; Acet, Ömür
Exosomes are microscopic vesicles secreted by cells, serving as carriers of diverse biological substances and playing an essential role in the communication between cells. When meticulously engineered, these small extracellular vesicles transform into highly effective delivery systems for therapeutic agents, enabling the targeted administration of active pharmaceutical ingredients to specific organs, tissues, and cells. Exosomes play an indispensable role in a myriad of biological processes, including intercellular communication, the regulation of gene expression, apoptosis, inflammation, immunity, as well as cell maturation and differentiation. The versatile role of exosomes is largely attributed to their intricate cargoes and composition, which encompasses nucleic acids, proteins, and lipids. In this review, we present a comprehensive overview of state-of-the-art characterization and isolation techniques used for the study of exosomes, especially for exosome-based biomedical applications. We will discuss the potential use of exosomes in personalized treatments, their interactions with other nanostructures focusing on the biomolecule corona, as well as the challenges and future expectations. In conclusion, this review provides evidence that we will witness extremely important functions and advances with innovative therapeutic and diagnostic applications of exosomes in the biomedical field.
