Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/395
Browse
8 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 Citation - WoS: 1Citation - Scopus: 1Prediction of Preference and Effect of Music on Preference: A Preliminary Study on Electroencephalography from Young Women(Tubitak Scientific & Technological Research Council Turkey, 2019-03-01) Yilmaz, Bulent; Gazeloglu, Cengiz; Altindis, FatihNeuromarketing is the application of the neuroscientific approaches to analyze and understand economically relevant behavior. In this study, the effect of loud and rhythmic music in a sample neuromarketing setup is investigated. The second aim was to develop an approach in the prediction of preference using only brain signals. In this work, 19-channel EEG signals were recorded and two experimental paradigms were implemented: no music/silence and rhythmic, loud music using a headphone, while viewing women shoes. For each 10-sec epoch, normalized power spectral density (PSD) of EEG data for six frequency bands was estimated using the Burg method. The effect of music was investigated by comparing the mean differences between music and no music groups using independent two-sample t-test. In the preference prediction part sequential forward selection, k-nearest neighbors (k-NN) and the support vector machines (SVM), and 5-fold cross-validation approaches were used. It is found that music did not affect like decision in any of the power bands, on the contrary, music affected dislike decisions for all bands with no exceptions. Furthermore, the accuracies obtained in preference prediction study were between 77.5 and 82.5% for k-NN and SVM techniques. The results of the study showed the feasibility of using EEG signals in the investigation of the music effect on purchasing behavior and the prediction of preference of an individual.Article Citation - Scopus: 6Network Intrusion Detection Based on Machine Learning Strategies: Performance Comparisons on Imbalanced Wired, Wireless, and Software-Defined Networking (SDN) Network Traffics(Turkiye Klinikleri, 2024-07-26) Hacilar, Hilal; Aydin, Zafer; Güngör, Vehbi ÇağrıThe rapid growth of computer networks emphasizes the urgency of addressing security issues. Organizations rely on network intrusion detection systems (NIDSs) to protect sensitive data from unauthorized access and theft. These systems analyze network traffic to detect suspicious activities, such as attempted breaches or cyberattacks. However, existing studies lack a thorough assessment of class imbalances and classification performance for different types of network intrusions: wired, wireless, and software-defined networking (SDN). This research aims to fill this gap by examining these networks’ imbalances, feature selection, and binary classification to enhance intrusion detection system efficiency. Various techniques such as SMOTE, ROS, ADASYN, and SMOTETomek are used to handle imbalanced datasets. Additionally, eXtreme Gradient Boosting (XGBoost) identifies key features, and an autoencoder (AE) assists in feature extraction for the classification task. The study evaluates datasets such as AWID, UNSW, and InSDN, yielding the best results with different numbers of selected features. Bayesian optimization fine-tunes parameters, and diverse machine learning algorithms (SVM, kNN, XGBoost, random forest, ensemble classifiers, and autoencoders) are employed. The optimal results, considering F1-measure, overall accuracy, detection rate, and false alarm rate, have been achieved for the UNSW-NB15, preprocessed AWID, and InSDN datasets, with values of [0.9356, 0.9289, 0.9328, 0.07597], [0.997, 0.9995, 0.9999, 0.0171], and [0.9998, 0.9996, 0.9998, 0.0012], respectively. These findings demonstrate that combining Bayesian optimization with oversampling techniques significantly enhances classification performance across wired, wireless, and SDN networks when compared to previous research conducted on these datasets. © 2024 Elsevier B.V., All rights reserved.Article Citation - Scopus: 25Lyapunov Exponent Enhancement in Chaotic Maps With Uniform Distribution Modulo One Transformation(Akif AKGUL, 2022-03-30) Ablay, GünyazMost of the chaotic maps are not suitable for chaos-based cryptosystems due to their narrow chaotic parameter range and lacking of strong unpredictability. This work presents a nonlinear transformation approach for Lyapunov exponent enhancement and robust chaotification in discrete-time chaotic systems for generating highly independent and uniformly distributed random chaotic sequences. The outcome of the new chaotic systems can directly be used in random number and random bit generators without any post-processing algorithms for various information technology applications. The proposed Lyapunov exponent enhancement based chaotic maps are analyzed with Lyapunov exponents, bifurcation diagrams, entropy, correlation and some other statistical tests. The results show that excellent random features can be accomplished even with one-dimensional chaotic maps with the proposed approach. © 2023 Elsevier B.V., All rights reserved.Article Citation - Scopus: 2Evaluating the Microbial Growth Kinetics and Artificial Gastric Digestion Survival of a Novel Pichia Kudriavzevii FOL-04(Field Crops Central Research Institute, 2022-06-15) Gumustop, Ismail; Ortakci, FatihPresent study aims to explore Pichia kudriavzevii FOL-04 (FOL-04)’s: i) survival against artificial gastric juice (AGJ) and artificial bile juice (ABJ), ii) growth kinetics in shake flask (SF) and fed-batch trials (FBT). Survival of FOL-04 as measured by relative cell density (RCD) against AGJ and ABJ was screened at four different pH-levels (control, 3, 2, 1.5) and ox-bile concentrations (control, 0.2%, 1%, 2%), respectively. Growth kinetics was calculated by periodic measurement of OD<inf>600</inf> in SF (225 rpm, 30°C) or in FBT using exponential feeding regimen where pH, dissolved-oxygen and temperature were controlled at 5.5, 21%, and 30°C, respectively. The doubling-time, maximum specific growth rate, and final cell densities achieved for SF and FBT were 81.7min, 1.67, 11.79 and 170.4 min, 4.75, 37.95, respectively. RCDs calculated were similar for pH=3 and control vs both were significantly higher(p<0.05) than pH=1.5 and 2 with the latter two pH-levels were not significantly different(p>0.05). RCDs were similar across control, 0.2%, and 1% ox-bile levels(p>0.05). However, 2% ox-bile yielded significantly lower RCD (p<0.05) compared to all except 1%. FOL-04 is a potential probiotic candidate showing robustness against AGJ and ABJ and remarkable biomass increase was achieved when grown under FBT which could pave the way for developing a yeast-based probiotic using this strain. © 2023 Elsevier B.V., All rights reserved.Article Citation - Scopus: 1Electricity Load Forecasting Using Deep Learning and Novel Hybrid Models(Sakarya University, 2022-02-28) Sutcu, Muhammed; Şahi̇n, Kübra Nur; Koloğlu, Yunus; Çelikel, Mevlüt Emirhan; Gulbahar, Ibrahim TümayLoad forecasting is an essential task which is executed by electricity retail companies. By predicting the demand accurately, companies can prevent waste of resources and blackouts. Load forecasting directly affect the financial of the company and the stability of the Turkish Electricity Market. This study is conducted with an electricity retail company, and main focus of the study is to build accurate models for load. Datasets with novel features are preprocessed, then deep learning models are built in order to achieve high accuracy for these problems. Furthermore, a novel method for solving regression problems with classification approach (discretization) is developed for this study. In order to obtain more robust model, an ensemble model is developed and the success of individual models are evaluated in comparison to each other. © 2025 Elsevier B.V., All rights reserved.Article Comparative Assessment of Smooth and Non-Smooth Optimization Solvers in Hanso Software(Ramazan Yaman, 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.