Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/395
Browse
61 results
Search Results
Article Fine-Tuning Large Language Models for Turkish Flutter Code Generation(Sakarya University, 2025-12-29) Uluırmak, Buğra Alperen; Kurban, RifatThe rapid advancement of large language models (LLMs) for code generation has largely centered on English programming queries. This paper focuses on a low-resource language scenario, specifically Turkish, in the context of Flutter mobile app development. Two representative LLMs (a 4B-parameter multilingual model and a 3B code-specialized model) on a new Turkish question-and-answer dataset for Flutter/Dart are fine-tuned in this study. Fine-tuning with parameter-efficient techniques yields dramatic improvements in code generation quality: Bilingual Evaluation Understudy (BLEU), Recall-Oriented Understudy for Gisting Evaluation (ROUGE-L), Metric for Evaluation of Translation with Explicit Ordering (METEOR), Bidirectional Encoder Representations from Transformers Score (BERTScore), and CodeBLEU scores show significant increases. The rate of correct solutions increased from ~30–70% (for base models) to 80–90% after fine-tuning. The performance trade-offs between models are analyzed, revealing that the multilingual model slightly outperforms the code-focused model in accuracy after fine-tuning. However, the code-focused model demonstrates faster inference speeds. These results demonstrate that even with very limited non-English training data, customizing LLMs can bridge the gap in code generation, enabling high-quality assistance for Turkish developers comparable to that for English. The dataset was released on GitHub to facilitate further research in multilingual code generation.Article Comparative Assessment of Smooth and Non-Smooth Optimization Solvers in HANSO Software(Balikesir University, 2021-10-27) Tor, Ali HakanThe aim of this study is to compare the performance of smooth and nonsmooth mization) software. The smooth optimization solver is the implementation of the Broyden-Fletcher-Goldfarb-Shanno (BFGS) method and the nonsmooth optimization solver is the Hybrid Algorithm for Nonsmooth Optimization. More precisely, the nonsmooth optimization algorithm is the combination of the BFGS and the Gradient Sampling Algorithm (GSA). We use well-known collection of academic test problems for nonsmooth optimization containing both convex and nonconvex problems. The motivation for this research is the importance of the comparative assessment of smooth optimization methods for solving nonsmooth optimization problems. This assessment will demonstrate how successful is the BFGS method for solving nonsmooth optimization problems in comparison with the nonsmooth optimization solver from HANSO. Performance profiles using the number iterations, the number of function evaluations and the number of subgradient evaluations are used to compare solvers.Article Deprem Anında ve Sonrasında Dönüşebilen Tasarım Önerileri: Günlük Kullanımdan Hayat Kurtaran Birimlere(Afet ve Acil Durum Yonetimi Baskanligi (AFAD), 2025-12-30) Özmen, Nihan Muş; Kurtuluş, Vacide BetülDepremler, yıkılan binalar ve devrilen eşyalar nedeniyle meydana gelen yaralanmalarla insan hayatını ciddi şekilde tehdit etmektedir. Bu çalışma, mobilyalara kendini kurtarma alanları entegre ederek ezilme yaralanmalarını önlemeyi ve deprem sonrası barınma ihtiyaçlarını karşılamayı amaçlayan tasarım önerileri sunmaktır. 2022 Güz döneminde Abdullah Gül Üniversitesi Deneysel Tasarım Stüdyosu kapsamında yürütülen çalışma, tanınmış bir mobilya tasarım firmasıyla iş birliği içinde gerçekleştirilmiştir. Öğrenciler hem günlük yaşamda işlevsel hem de afet anlarında acil barınma alanı olarak kullanılabilecek çift amaçlı mobilya tasarımları geliştirmiştir. Projeler, deprem sırasında ve sonrasında kullanılmak üzere iki ana kategoriye ayrılmış; yaşam ve çalışma mekânlarında, dönüşebilir mobilya tasarımlarına odaklanmıştır. Bulgular, işlevsel ve uyarlanabilir tasarımlar yoluyla risk azaltma ve deprem sonrası uyum konularında yenilikçi yaklaşımları ortaya koymaktadır.Article Noninvasive Condition Monitoring for Eccentricity Fault Detection in Large Hydro Generators(TÜBİTAK Scientific & Technological Research Council Turkey, 2026-01-16) Lemeski, Atena Tazikeh; Tekgun, Didem; Keysan, Ozan; Leblebicioglu, Kemal; Gol, Murat; Leblebicioglu, Mehmet KemalEccentricity faults in electric machines remain a critical concern, as they generate uneven magnetic forces that increase vibration and noise, ultimately raising the risk of premature motor failure. This study proposes a method for the early detection of dynamic eccentricity (DE) faults in hydropower plants through an advanced optimization-based parameter identification technique integrated with finite element analysis (FEA). Finite element modeling (FEM) is first used to analyze an existing salient-pole synchronous generator (SPSG) from a hydroelectric power plant in T & uuml;rkiye. The effects of DE faults on the SPSG's magnetic equivalent circuit parameters are then examined under various fault severities. A comprehensive hydropower plant model-including the synchronous generator, governor, and excitation system-is developed in MATLAB/Simulink, with all input parameters obtained from real plant data and equivalent circuit variations extracted from FEA. After completing the modeling stage, including fault scenarios, MATLAB and Simulink are employed together to estimate key magnetic equivalent circuit parameters using a modified particle swarm optimization (MPSO) algorithm, achieving highly accurate parameter estimation. Since the hydropower system allows measurement of the three-phase output currents, parameter estimation is performed based on current variations under different fault conditions. The simulation results verify the method's ability to detect faults with high accuracy; thus, this integrated and noninvasive approach provides a robust framework for ensuring the operational reliability and longevity of large hydro generators.Article Citation - Scopus: 5University Librarians’ Perceptions Of Artificial Intelligence, Its Application Areas İn Libraries, And The Future(University and Research Librarians Association (UNAK), 2024-12-26) Cuhadar, S.; Mert, S.; Gezer, Ç.; Helvacioğlu, E.; Arus, O.; Aslan, Ö.; Atli, S.; Gurdal, Gultekin; Erken, MehmetToday, libraries are among the institutions affected by changing technology and innovations. The popularization of artificial intelligence (AI) technologies has also begun to transform library services. In this research, a survey was conducted to determine the adjustments that university libraries in Turkey have made and plan to make during the development process of AI technologies and applications, and to identify the services they have developed specific to the relevant period. The survey was carried out with the participation of 111 university library managers from 208 university libraries in Turkey. Through the analysis of the data, the status, knowledge, and awareness levels of university libraries regarding AI technologies and applications were determined, and measures and recommendations were presented to improve deficiencies and weaknesses. This research is the first and most comprehensive study conducted in Turkey by obtaining opinions and suggestions from university library managers on artificial intelligence. The research findings revealed that university libraries use AI applications such as ChatGPT, Gemini, and Grammarly to a certain extent; however, they have needs in developing institutional policies, enhancing personnel competencies, and planning related to AI. © 2024 University and Research Librarians Association (UNAK). All rights reserved.Article Spatial Dimension of the Local Phenomenon in Kayseri(Gazi University, Faculty of Engineering Architecture, 2025-12-31) Ozmen, Nihan Mus; Asiliskender, BurakKayseri is in the centre of Anatolia, at the intersection of trade and military routes, and possesses a rich cultural heritage. Throughout its history, the city has hosted various civilizations, developing around a central castle and continuing to expand, particularly after the 19th century. Kayseri has long served as a meeting point for diverse cultures. Within this diversity, families known as locals, whose origins date back to the oldest neighbourhoods within the city walls, have held significant mercantile power. These local families regard themselves as the actual owners of Kayseri and have influenced the city's developmental trajectory. Over time, they have moved outward from the centre to newly developed neighbourhoods, first to the north and then to the east. This study examines the urban development of Kayseri in the 20th century and the spatial mobility of these local families. It employs qualitative methods such as ethnographic observation, oral history interviews, and GIS-based thematic mapping to analyse these movements in a multi-layered way. The study also aims to understand Kayseri's socio-cultural dynamics and historical texture by investigating the role of local families in the city's physical and functional transformations. In this context, it addresses the physical and functional changes in neighbourhoods vacated by these relocations.Article Modeling and Simulation of Dynamic Energy Management Systems for Smart Buildings(TÜBİTAK, 2025-11-25) Ozel, O.; Rıfat Boynueğrİ, A.; Yigit, H.; Tekgun, B.; Boynuegri, Ali RifatThis study presents a dynamic energy management system tailored for smart residential buildings, integrating thermal and electrical models to achieve both natural gas and electricity bill cost reduction. By harnessing wind and solar energy sources, the system aims to meet the diverse energy needs of modern homes. Through load shifting and thermal storage strategies, known as power-to-heat (P2H) approaches, the system ensures efficient renewable energy utilization while maintaining resident comfort. Validation of the proposed system was conducted using real-world data from the Yıldız Technical University Smart Home Laboratory, demonstrating its practical applicability and effectiveness. Results indicate significant reductions in both natural gas and electricity consumption, leading to substantial cost savings. Specifically, the proposed system reduced natural gas consumption by 3.79% and electricity consumption by 35.62%, highlighting its potential to enhance energy efficiency and sustainability in residential settings. © This work is licensed under a Creative Commons Attribution 4.0 International License.Article Developing a Label Propagation Approach for Cancer Subtype Classification Problem(TUBITAK, 2021) Güner, P.; Bakir-Güngör, B.; Coşkun, M.; Şahan, Pınar GünerCancer is a disease in which abnormal cells grow uncontrollably and invade other tissues. Several types of cancer have various subtypes with different clinical and biological implications. Based on these differences, treatment methods need to be customized. The identification of distinct cancer subtypes is an important problem in bioinformatics, since it can guide future precision medicine applications. In order to design targeted treatments, bioinformatics methods attempt to discover common molecular pathology of different cancer subtypes. Along this line, several computational methods have been proposed to discover cancer subtypes or to stratify cancer into informative subtypes. However, existing works do not consider the sparseness of data (genes having low degrees) and result in an ill-conditioned solution. To address this shortcoming, in this paper, we propose an alternative unsupervised method to stratify cancer patients into subtypes using applied numerical algebra techniques. More specifically, we applied a label propagation-based approach to stratify somatic mutation profiles of colon, head and neck, uterine, bladder, and breast tumors. We evaluated the performance of our method by comparing it to the baseline methods. Extensive experiments demonstrate that our approach highly renders tumor classification tasks by largely outperforming the state-of-the-art unsupervised and supervised approaches. © 2022 Elsevier B.V., All rights reserved.Article Citation - WoS: 1Comprehensive Prediction of FBN1 Targeting Mirnas: A Systems Biology Approach for Marfan Syndrome(Galenos Publishing House, 2025-09-22) Orhan, M.E.; Demirci, Y.M.; Saçar Demirci, M.D.S.; Demirci, Muserref Duygu SacarObjective: Marfan syndrome (MFS) is a genetic connective tissue disorder primarily caused by mutations in the FBN1 gene. Emerging evidence highlights the regulatory role of microRNAs (miRNAs) in modulating gene expression in MFS, but a systematic investigation into miRNAs targeting FBN1 is lacking. This study aimed to comprehensively identify miRNAs interacting with the FBN1 transcript to reveal potential molecular regulators and therapeutic targets. Methods: Human miRNA sequences were retrieved from miRBase (Release 22.1), and the canonical FBN1 transcript (RefSeq: NM_000138.5) was used for target prediction. Computational interaction analysis was conducted using the psRNATarget server with stringent parameters to detect potential miRNA binding sites. Expression profiles and disease associations of the top candidate miRNAs were further investigated through database integration and literature review. Results: Out of 2656 human mature miRNAs analyzed, 251 were predicted to bind FBN1, with the hsa-miR-181 family exhibiting the highest number of predicted interactions. Evidence from the literature highlighted dysregulation of hsa-miR-181 expression in MFS patients, suggesting a functional role in disease pathophysiology. Conclusion: This study identifies key members of the hsa-miR-181 family as post-transcriptional regulators of FBN1, offering new insights into miRNA-driven mechanisms in MFS. These findings support the potential of RNA-based diagnostics and therapeutic strategies targeting miRNA-FBN1 interactions. ©Copyright 2025 The Author.Article Forecasting the Consumer Price Index in Türkiye Using Machine Learning Models: A Comparative Analysis(Gazi Univ, 2025-09-01) Söylemez, İsmet; Ünlü, Ramazan; Nalici, Mehmet ErenThis study utilizes machine learning models to forecast Türkiye's Consumer Price Index (CPI), thereby addressing a critical gap in inflation prediction methodologies. The central research problem involves the forecasting of CPI in a volatile economic environment, which is essential for informed policymaking. The primary objective of this study is to evaluate the performance of three machine learning models, such as Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM), in forecasting CPI over periods ranging from one to six months, utilizing data from 2012 to 2024. The study's unique contribution lies in the application of the \"SelectKBest\" method, which identifies the most relevant indices, thereby enhancing the efficiency of the models. An ensemble method, Averaging Voting, is also employed to combine the strengths of these models, producing more accurate and robust predictions. The findings indicate that while the RF model consistently generates the most accurate forecasts across all shifts, the SVM model demonstrates a particular strength in the domain of short-term predictions. The ensemble model demonstrates a substantial performance improvement, with a R2 value of 0.962 for one-month ahead of estimates and 0.956 for five-month forecasts. This combined approach has been shown to outperform individual models, offering a more reliable framework for CPI forecasting. The findings offer valuable insights for economic policymakers, enabling more precise and stable inflation predictions in Türkiye.