Scopus İndeksli Yayınlar Koleksiyonu

Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/8651

Browse

Filters

20257

Settings

Publisher: search.filters.publisher.IEEEWoS Q: search.filters.wosQuartile.N/A

Search Results

Now showing 1 - 7 of 7
  • Conference Object
    Citation - WoS: 1
    Citation - Scopus: 1
    Dengesiz Epilepsi Veri Seti İçin Sınıflandırmada Farklı SMOTE Yöntemlerinin Etkileri
    (Institute of Electrical and Electronics Engineers Inc., 2025) Calis, Ahmet Gokay; Ergezer, Halit
    In this study, the effects of different SMOTE methods on machine learning algorithms for the imbalanced epilepsy dataset were investigated. After filtering, the imbalanced dataset was balanced with 5 different SMOTE methods and classified with various machine learning algorithms. Coarse-K-Nearest Neighbor, Bagged Trees, and Artificial Neural Networks models were evaluated in epilepsy detection. The performance of these different models was compared with Matthews Correlation Coefficient (MCC) and F1 Score metrics. The results showed that the Borderline-SMOTE algorithm had the highest F1 Score and MCC values among all machine learning algorithms. © 2025 Elsevier B.V., All rights reserved.
  • Conference Object
    AviBERT: Transformer Tabanlı Hava Aracı Metni Sınıflandırma
    (Institute of Electrical and Electronics Engineers Inc., 2025) Unal, Muhammed Cihat; Yurtalan, Gokhan; Karatas, Yahya Bahadir; Karamanlioglu, Alper; Demirel, Berkan
    In recent years, transformer-based models pre-trained on extensive corpora have played a critical role in the advancement of Natural Language Processing methodologies. Particularly, methods based on BERT have demonstrated remarkable performance across various tasks by offering robust capabilities in deeply understanding texts semantically. However, despite these advancements, there is a notable scarcity of studies applying these technologies in the aviation sector. This paper develops a multi-class classification model for aviation-specific texts using variants of BERT. The study encompasses the processes of collecting web content related to aircraft, labeling and model training. The details of the dataset are explained and the outcomes of the study are assessed based on the macro F1-score and accuracy of different models. © 2025 Elsevier B.V., All rights reserved.
  • Conference Object
    Model Enhancement for UAV Stealth in X-Band
    (Institute of Electrical and Electronics Engineers Inc., 2025) Unalir, Dizdar; Yalcinkaya, Bengisu; Aydin, Elif
    With the rapid advancement of technology, radar detection techniques continue to evolve, challenging the effectiveness of traditional unmanned aerial vehicles (UAVs) stealth techniques. As the usage of UAVs in military applications expands, the need for effective radar cross section reduction (RCSR) methods to enhance their stealth capabilities has grown significantly. In this study, we propose an enhancement of a previously developed Low-RCS UAV model, focusing on RCSR with shaping technique in the X-band. For the identification and optimization of the UAV model’s highly reflective components, a detailed simulative analysis of the RCS was performed using CST Studio Suite Environment. The modifications are applied to the body and leg components to minimize radar reflections. Simulation results demonstrated that the proposed enhancements significantly reduced RCS values compared to the original Low-RCS UAV model. A total of 13 dBsm reduction in RCS was observed compared to the traditional UAV models. Comparative analysis for different frequencies in X-Band and various aspect angles confirmed the effectiveness of the improved design, validating its potential for stealth applications. The findings can contribute to the research in UAV stealth technology and provide insights into future low-visibility UAV designs. © 2025 Elsevier B.V., All rights reserved.
  • Conference Object
    Unified Lf-Norm Robust Fitting for Linear Models
    (Institute of Electrical and Electronics Engineers Inc., 2025) Nar, Fatih; Saran, Murat; Saran, Ayse Nurdan; Sen, Baha
    In statistical learning, accurately estimating model parameters is crucial for reliable predictions. Managing residuals, the differences between observed and predicted values, is a key challenge. In regression, the residual penalty choice strongly affects model performance. The L<inf>2</inf>-norm penalty aligns with the least-squares approach, while the L<inf>1</inf>-norm provides robust fitting by minimizing the influence of outliers. To generalize models, the weights can be regularized using either the L<inf>2</inf>-norm or L<inf>1</inf>-norm, corresponding to Ridge and LASSO regularization, respectively. Many methods have been developed to penalize residuals and model weights, resulting in diverse cost functions optimized by specific numerical solvers. In this study, we propose the smooth L<inf>f</inf>-norm, a quasi-norm, as a unified framework for penalizing both residuals and model weights in linear models. Our efficient and robust numerical minimization scheme ensures fast and accurate fitting by minimizing our novel cost function. © 2025 Elsevier B.V., All rights reserved.
  • Conference Object
    Optimal Fixed-Wing UAV Rendezvous Via LQR-Based Longitudinal Control
    (IEEE, 2025) Buyukekiz, Kadir Bulathan; Ergezer, Halit
    This paper proposes an optimal control-based rendezvous strategy for fixed-wing Unmanned Aerial Vehicles (UAVs) using a Linear Quadratic Regulator (LQR). The goal is precisely tracking a moving target while maintaining flight stability and avoiding predefined restricted areas. The controller optimally adjusts UAVs flight parameters to minimize trajectory errors and enhance robustness against environmental disturbances. A penalty-based method is integrated to prevent UAVs from entering restricted areas while ensuring smooth trajectory adaptation. The proposed approach has been tested in MATLAB simulations under multiple scenarios, demonstrating its effectiveness in achieving stable and efficient rendezvous maneuvers. The results confirm that LQR-based control and adaptive penalty mechanisms offer a practical solution for fixed-wing UAV operations in constrained environments.
  • Conference Object
    Controller Synthesis With a Cone Complementarity Linearization Algorithm for CACC System Under Time-Varying Delay
    (IEEE, 2025) Bingol, Hilal
    Cooperative Adaptive Cruise Control (CACC) is an intelligent vehicle technology that enables vehicle follow-up in a small inter-vehicle distance. Stability is the main property of the CACC system that prevents any signal fluctuations throughout the vehicle string. In CACC system, time delay has a negative impact on string stability. The reason is that constant and varying time delays are unavoidable in real traffic. Here, controller gains should be synthesized under constant and time-varying communication delays to satisfy L-2-string stability conditions (in the Lyapunov sense). Contrary to previously studied convex method, controller gains are now synthesized with iterative nonlinear minimization algorithm and Cone Complementarity Linearization (CCL) method. The results show that the new CCL approach provides more accurate and practical stability bounds. Hence, this shows the potential of CCL to perform more sensitive analyses. The results obtained are evaluated by simulations with the heterogeneous CACC system.
  • Conference Object
    Topic-Aware Multi-Class Classification for Financial Complaints: Comparing BERTopic With Classical Machine Learning Algorithms
    (Institute of Electrical and Electronics Engineers Inc., 2025) Uguz, Sezer; Kumbasar, Mert; Tokdemir, Gul
    In today's digital world, customers can utilize a variety of communication channels, such as business emails, consumer forms, feedback platforms, and dedicated complaint websites, to communicate their complaints. This study compares the performance of the supervised Bidirectional Encoder Representations from Transformers for Topic Modeling (BERTopic) with traditional machine learning algorithms, including Random Forest (RF), Support Vector Machines (SVM), Logistic Regression (LR), Naive Bayes (NB), K-nearest Neighbors (KNN), and eXtreme Gradient Boosting (XGBoost), for multi-class classification of financial customer complaints. The dataset consists of 16,715 balanced training data and 3,808 test data across five different categories, with the financial complaint data. Experimental results demonstrate that traditional machine learning models, particularly XGBoost, SVM, and LR, achieved the highest classification performance with accuracy rates close to 88%. BERTopic showed a competitive performance with an accuracy of 82.48%. The results suggest that while BERTopic offers interpretability advantages through topic modeling techniques, traditional algorithms provide higher accuracy. This study highlights the promising potential for future financial text analysis and customer complaint classification using hybrid methods, which could lead to more detailed, topic-aware classification approaches. © 2025 IEEE.