WoS İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12573/394
Browse
13 results
Search Results
Article An Attention-Based Autoencoder Model with Gated Recurrent Unit for Stock Price Movement Prediction(Springer Nature, 2026) Kolukisa, Burak; Bakir-Gungor, Burcu; Akkas, HuseyinPredicting stock movements is crucial for investors looking to maximize their profits in rapidly changing financial markets. However, noise in stock price data makes it harder to detect the trends and ultimately decreases the performance of predictive models. To address these challenges that are caused by noise, this study proposes two novel models, i.e., an Attention-based Autoencoder (ABA) and an Attention-based Variational Autoencoder with Gated Recurrent Units (AGRUA), which uniquely integrate denoising, attention, and GRU layers. Firstly, a data set is created by adding 9 different technical indicators to the historical data of Borsa Istanbul (BIST 30) stocks. Secondly, proposed methods and other well-known deep learning models were used to remove noise from the data sets. Finally, each denoised dataset was fed separately to the Extreme Gradient Boosting model and subjected to a buy-sell process. The results were measured using trading indicators such as amount of the profits and Sharpe Ratio, Sortino Ratio and Maximum Drawdown. The proposed models produced substantial financial gains, with AGRUA achieving the highest total profit and ABA achieving the lowest average Maximum Drawdown, thereby demonstrating superior risk-adjusted performance. Lastly, Friedman and Nemenyi tests confirmed that AGRUA and ABA surpass most of the benchmarks in profit and risk-adjusted returns. The performance of the proposed method demonstrates its value in capturing nonlinear market patterns and improving decision accuracy, emphasizing the need for noise reduction before forecasting.Article Frequency-Based Deep Occlusion Awareness Instance Segmentation(MDPI, 2026-02-26) Guzel, Yasin; Aydin, Zafer; Talu, Muhammed FatihOne major challenge faced by deep learning-based methods that detect target objects in the form of bounding boxes is object occlusion. High degrees of occlusion significantly diminish the accuracy of instance segmentation. Nonetheless, complex-valued Fourier descriptors can robustly represent object boundaries using minimal information. In this study, the impact of integrating Fourier descriptors-renowned for their strong representational capacity-with deep network models (UNet) that exhibit high generalization performance on instance segmentation accuracy was investigated. Within the scope of the research, nine network models were designed based on different strategies for utilizing frequency components. These variants fall into four strategy families: (i) UNet-style spectrum regression on fixed low-frequency windows (FUNet), (ii) magnitude-guided frequency selection/ROI construction (FUNet-Thr, FUNet-BBox), (iii) sequence models over tokenized FFT coefficients (BiLSTM Patch/Sorted), and (iv) encoder-only spectrum predictors with different depth/capacity (EncoderFFT1/2). To fairly evaluate the models' performance in segmenting objects subjected to disruptive factors (e.g., occlusion, blurring, noise), a specialized synthetic dataset was prepared. The task is formulated as single-target (single-instance), single-class segmentation. This dataset, automatically generated according to initial parameter values, contains images of objects moving at various speeds within a single frame. Among these models, the one termed FUNet, which relies on partial matching of central frequency components, achieved the highest segmentation accuracy despite the disruptive effects. Under the challenging Dataset 8 setting, the proposed FUNet achieved the highest overlap-based performance (Dice = 0.9329, IoU = 0.8842) among Attention U-Net, U-Net, and FourierNet, with statistically significant gains confirmed by paired per-image tests.Article Non-Contact Acoustic Screening for Sleep Apnea: A Subject-Aware Deep Learning Approach(Springer Science and Business Media Deutschland GmbH, 2026-02-11) Aygün Çakıroğlu, M.; Kizilkaya Aydoǧan, E.; Bolatturk, Ö.F.; Aydoğan, S.; Ismailoǧullari, S.; Delice, Y.Purpose: To explore the feasibility of using camera-derived, non-contact audio synchronized with PSG for clinically relevant sleep-apnea classification, and to benchmark compact deep models under a subject-aware design using a previously unstudied, real-world dataset. Methods: Thirty-two adults underwent simultaneous polysomnography (PSG) and camera-based non-contact audio recording. The synchronized audio segments were used to train and compare three compact deep-learning architectures (convolutional, attention-augmented, and transformer-based) under a subject-aware evaluation design that prevented identity leakage. Model performance and calibration were assessed at both segment and subject levels using standard statistical tests. Results: Subject-level evaluation was based on a very small, imbalanced test set of six subjects (one positive). Within this limited yet previously unstudied local dataset, the CNN_trans model achieved an apparent perfect ranking performance (AUC = 1.00; 95% CI 0.00–1.00), though this likely reflects the small, imbalanced test cohort, with recall = 1.00 and precision = 0.55. The wide confidence interval reflects substantial statistical uncertainty, and DeLong comparisons showed no significant AUC difference between CNN_trans and CNN_att (ΔAUC = − 0.042; p = 0.43). Conclusion: PSG-synchronized, non-contact audio supports accurate and well-calibrated sleep-apnea classification with compact deep models. This subject-aware evaluation suggests that contactless acoustic monitoring may have potential clinical relevance, motivating larger, multi-site validation. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2026.Article Deep-Learning Detection of Open-Apex Teeth on Panoramic Radiographs Using YOLO Models(Springer, 2025-12-23) Edik, Merve; Celebi, Fatma; Cukurluoglu, AykaganObjectivesThe use of deep learning in detecting teeth with open apices can prevent the need for additional radiographs for patients. The presented study aims to detect open-apex teeth using You Only Look Once (YOLO)-based deep learning models and compare these models.MethodsA total of 966 panoramic radiographs were included in the study. Open-apex teeth in panoramic radiographs were labeled. During the labeling process, they were divided into 6 classes in the maxilla and mandible, namely incisors, premolars, and molars. AI models YOLOv3, YOLOv4, and YOLOv5 were used. To evaluate the performance of the three detection models, both overall and separately for each class in the test dataset, precision, recall, average precision (mAP), and F1 score were calculated.ResultsYOLOv4 achieved the highest overall performance with a mean average precision (mAP) of 87.84% at IoU (Intersection over Union) 0.5 (mAP@0.5), followed by YOLOv5 with 85.6%, and YOLOv3 with 84.46%. Regarding recall, YOLOv4 also led with 90%, while both YOLOv3 and YOLOv5 reached 89%. Moreover, the F1 score was the highest for YOLOv4 (0.87), followed by YOLOv3 (0.86) and YOLOv5 (0.85).ConclusionsIn this study, YOLOv3, YOLOv4, and YOLOv5 were evaluated for the detection of open-apex teeth, and their mAP, recall, and F1 scores exceeded 84%. Deep learning-based systems can provide faster and more accurate results in the detection of open-apex teeth. This may help reduce the need for additional radiographs from patients and aid dentists by saving time.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 Network Intrusion Detection Based on Machine Learning Strategies: Performance Comparisons on Imbalanced Wired, Wireless, and Software Defined Networking (SDN) Network Traffics(Tubitak Scientific & Technological Research Council Turkey, 2024) Hacilar, Hilal; Aydin, Zafer; Gungor, Vehbi CagriThe 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 autoenco der (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.Article Citation - WoS: 13Citation - Scopus: 21IGPRED: Combination of Convolutional Neural and Graph Convolutional Networks for Protein Secondary Structure Prediction(Wiley, 2021-05-25) Gormez, Yasin; Sabzekar, Mostafa; Aydin, ZaferThere is a close relationship between the tertiary structure and the function of a protein. One of the important steps to determine the tertiary structure is protein secondary structure prediction (PSSP). For this reason, predicting secondary structure with higher accuracy will give valuable information about the tertiary structure. Recently, deep learning techniques have obtained promising improvements in several machine learning applications including PSSP. In this article, a novel deep learning model, based on convolutional neural network and graph convolutional network is proposed. PSIBLAST PSSM, HHMAKE PSSM, physico-chemical properties of amino acids are combined with structural profiles to generate a rich feature set. Furthermore, the hyper-parameters of the proposed network are optimized using Bayesian optimization. The proposed model IGPRED obtained 89.19%, 86.34%, 87.87%, 85.76%, and 86.54% Q3 accuracies for CullPDB, EVAset, CASP10, CASP11, and CASP12 datasets, respectively.Article Citation - WoS: 4Citation - Scopus: 7IGPRED-Multitask: A Deep Learning Model to Predict Protein Secondary Structure, Torsion Angles and Solvent Accessibility(IEEE Computer Soc, 2023-03-01) Gormez, Yasin; Aydin, ZaferProtein secondary structure, solvent accessibility and torsion angle predictions are preliminary steps to predict 3D structure of a protein. Deep learning approaches have achieved significant improvements in predicting various features of protein structure. In this study, IGPRED-Multitask, a deep learning model with multi task learning architecture based on deep inception network, graph convolutional network and a bidirectional long short-term memory is proposed. Moreover, hyper-parameters of the model are fine-tuned using Bayesian optimization, which is faster and more effective than grid search. The same benchmark test data sets as in the OPUS-TASS paper including TEST2016, TEST2018, CASP12, CASP13, CASPFM, HARD68, CAMEO93, CAMEO93_HARD, as well as the train and validation sets, are used for fair comparison with the literature. Statistically significant improvements are observed in secondary structure prediction on 4 datasets, in phi angle prediction on 2 datasets and in psi angel prediction on 3 datasets compared to the state-of-the-art methods. For solvent accessibility prediction, TEST2016 and TEST2018 datasets are used only to assess the performance of the proposed model.Article Citation - WoS: 2Citation - Scopus: 9Human Identification Using Palm Print Images Based on Deep Learning Methods and Gray Wolf Optimization Algorithm(Springer London Ltd, 2023-10-24) Alshakree, Firas; Akbas, Ayhan; Rahebi, JavadPalm print identification is a biometric technique that relies on the distinctive characteristics of a person's palm print to distinguish and authenticate their identity. The unique pattern of ridges, lines, and other features present on the palm allows for the identification of an individual. The ridges and lines on the palm are formed during embryonic development and remain relatively unchanged throughout a person's lifetime, making palm prints an ideal candidate for biometric identification. Using deep learning networks, such as GoogLeNet, SqueezeNet, and AlexNet combined with gray wolf optimization, we achieved to extract and analyze the unique features of a person's palm print to create a digital representation that can be used for identification purposes with a high degree of accuracy. To this end, two well-known datasets, the Hong Kong Polytechnic University dataset and the Tongji Contactless dataset, were used for testing and evaluation. The recognition rate of the proposed method was compared with other existing methods such as principal component analysis, including local binary pattern and Laplacian of Gaussian-Gabor transform. The results demonstrate that the proposed method outperforms other methods with a recognition rate of 96.72%. These findings show that the combination of deep learning and gray wolf optimization can effectively improve the accuracy of human identification using palm print images.Article Citation - WoS: 3Citation - Scopus: 2Enhancing Diagnostic Quality in Panoramic Radiography: A Comparative Evaluation of GAN Models for Image Restoration(Wiley, 2025-09-04) Kolukisa, Burak; Celebi, Fatma; Ersu, Nihal; Yucel, Kemal Selcuk; Canger, Emin Murat; Murat Canger, EminPanoramic imaging is a widely utilized technique to capture a comprehensive view of the maxillary and mandibular dental arches and supporting facial structures. This study evaluates the potential of the Generative Adversarial Network (GAN) models-Pix2Pix, CycleGAN, and RegGAN-in enhancing diagnostic quality by addressing combinations of common image distortions. A panoramic radiograph data set was processed to simulate four types of distortions: (i) blurriness, (ii) noise, (iii) combined blurriness and noise, and (iv) anterior-region-specific blurriness. Three GAN models were trained and analyzed using quantitative metrics such as the peak signal-to-noise ratio (PSNR) and the structural similarity index measure (SSIM). In addition, two oral and maxillofacial radiologists conducted qualitative reviews to assess the diagnostic reliability of the generated images. Pix2Pix consistently outperformed CycleGAN and RegGAN, achieving the highest PSNR and SSIM values across all types of distortions. Expert evaluations also favored Pix2Pix, highlighting its ability to restore image accuracy and enhance clinical utility. CycleGAN showed moderate improvements in noise-affected images but struggled with combined distortions, while RegGAN yielded negligible enhancements. These findings underscore its potential for clinical application in refining radiographic imaging. Future research should focus on combining GAN techniques and utilizing larger datasets to develop universally robust image enhancement models.