Browsing by Author "Ozkurt, Nalan"

Now showing 1 - 20 of 25

Citation - WoS: 1
Citation - Scopus: 1
24-hour Electricity Consumption Forecasting for Day Ahead Market with Long Short Term Memory Deep Learning Model
(IEEE, 2020) Nalan Ozkurt; Hacer Sekerci Oztura; Cuneyt Guzelis; Guzelis, Cuneyt; Oztura, Hacer Sekerci; Ozkurt, Nalan
In 2015 with the foundation of Energy Market Management Inc. AS (EPIAS) the production and pricing of electrical energy began to be made according to consumption estimates. In this study twenty-four hours energy consumption forecasting was made by using long short-term memory method and data was downloded from EPIAS's official web page for the Day Ahead Market. The data set used covers 1500 days between June 2016 and July 2020. The results obtained have been compared with EPIAS's own estimates and actual consumption data.
A real-time accurate positioning system using BLE and wireless mesh network in a shipyard environment
(2022) Hilal KILINÇ; EKREM ÖZGÜRBÜZ; NALAN OZKURT; HASAN HUSEYIN ERKAN; Kılınç, Hilal; Özgürbüz, Ekrem; Ozkurt, Nalan; Erkan, Hasan Huseyın
Digitalization of the shipyard environment is a challenging problem and also it is necessary for competing the international ship manufacturers. Thus this study introduced a real-time accurate positioning system that is an indispensable part of a digital manufacturing system. The system implementation and measurements took place in Sedef Shipyard the largest shipbuilding facility in Turkey. Since the shipyard includes indoor and outdoor environments Bluetooth Low Energy (BLE) systems provide the best solution for locating the staff. The most challenging problem is to determine the positions in the metallic surroundings. The constructed system solves this problem by placing gateways and sensors at essential locations and using a mesh network. With the designed user interface the position of the staff can be monitored accurately in real time and reports can be generated.
Citation - WoS: 15
Citation - Scopus: 23
A review of ADHD detection studies with machine learning methods using rsfMRI data
(John Wiley and Sons Ltd, 2024) Gürcan Taşpinar; Nalan Ǒzkurt; Taspinar, Gurcan; Ozkurt, Nalan
Attention deficit hyperactivity disorder (ADHD) is a common mental health condition that significantly affects school-age children causing difficulties with learning and daily functioning. Early identification is crucial and reliable and objective diagnostic tools are necessary. However current clinical evaluations of behavioral symptoms can be inconsistent and subjective. Functional magnetic resonance imaging (fMRI) is a non-invasive technique that has proven effective in detecting brain abnormalities in individuals with ADHD. Recent studies have shown promising outcomes in using resting state fMRI (rsfMRI)-based brain functional networks to diagnose various brain disorders including ADHD. Several review papers have examined the detection of other diseases using fMRI data and machine learning or deep learning methods. However no review paper has specifically addressed ADHD. Therefore this study aims to contribute to the literature by reviewing the use of rsfMRI data and machine learning methods for detection of ADHD. The study provides general information about fMRI databases and detailed knowledge of the ADHD-200 database which is commonly used for ADHD detection. It also emphasizes the importance of examining all stages of the process including network and atlas selection feature extraction and feature selection before the classification stage. The study compares the performance advantages and disadvantages of previous studies in detail. This comprehensive approach may be a useful starting point for new researchers in this area. © 2024 Elsevier B.V. All rights reserved.
Analysis and Classification of Air Disc Brake Sounds in Time and Frequency Domain
(IEEE, 2018) Zeynep Ertekin; Nalan Ozkurt; Cem Yilmaz; Yilmaz, Cem; Ertekin, Zeynep; Ozkurt, Nalan
In this study analysis and classification of audio data collected from faulty air disc brakes has been carried out by Fourier Transform (FT). The sound data have been recorded by 2 identical Norsonic Type 1228 microphones in the laboratory of Ege Fren A.S. on a vehicle. The recorded data set which has been transferred into computer by a data acquisition board has been analyzed in Matlab. Number of zero crossings, mean variance entropy and spectral rolloff of Fourier coefficients have been used as features in order to distinguish normal and faulty brakes. These features have been classified with 10x10 crossvalidation by using kth nearest neighbour algorithm with a success rate of 96.23 %
Citation - Scopus: 3
Analysis of Cardiac Beats using Higher Order Spectra
(IEEE, 2014) Ibrahim Abdullahi Karaye; Sani Saminu; Nalan Ozkurt; Karaye, Ibrahim Abdullahi; Saminu, Sani; Ozkurt, Nalan; S Misra; C Ayo; N Omoregbe; B Odusote; A Adewumi
For early diagnosis of the heart failures the electrocardiography (ECG) is the most common method because of its simplicity and cost. Computer based analysis of ECG provides reliable and efficient tools in diagnostics of arrhythmias. With this objective there are lots of studies on automatic and semi-automatic ECG analysis. Like many biosignals ECG signals are nonlinear in nature higher order spectral analysis (HOS) is known to be a very good tool for the analysis of nonlinear systems producing good noise immunity. Thus in this study HOS analysis of ECG signals of normal heart rate right bundle branch block paced beat left bundle block branch and atrial premature beats have been studied in order to reveal the complex dynamics of ECG signals using the tools of nonlinear systems theory. Some of the general characteristics for each of these classes in the bispectrum and bicoherence plot for visual observation have been presented. For the extraction of R-R intervals well known Pan-Tompkins algorithm has been used and three higher order statistical parameters of skewness kurtosis and variance from these features have been computed. These features with statistical parameters fed into artificial neural network classifier (ANN) and obtained an average accuracy of 94.9%.
Citation - Scopus: 2
Arrhythmia Detection with Custom Designed Wavelet-based Convolutional Autoencoder
(Institute of Electrical and Electronics Engineers Inc., 2023) Oyku Eravci; Nalan Ǒzkurt; Eravci, Oyku; Ozkurt, Nalan; T. Yildirim , R. Chbeir , L. Bellatreche , C. Badica
In this study we provide a deep learning approach for the classification of arrhythmias that uses a customized wavelet-based autoencoder (AE) model for feature extraction and an MLP model for beat classification. The main objective of this paper is to evaluate the performance of auto-encoder based deep learning algorithms and to automatically classify five types of cardiac arrhythmias such as normal heartbeat (NSR) right bundle branch block (RBBB) left bundle branch block (LBBB) and atrial premature beats (APC) premature ventricular beats (PVC). The proposed approach received average scores of 99.8% 99.8% and 99.7% 99.7% for accuracy precision recall and F1 on publicly available datasets. The outcomes of the experiments demonstrate the significance of using deep learning based models in the diagnosis of cardiac disease. © 2023 Elsevier B.V. All rights reserved.
Citation - Scopus: 3
Beef Quality Assesment with Electronic Nose Based on an Application Specific Convolution Neural Network
(Institute of Electrical and Electronics Engineers Inc., 2021) Parvin Bulucu; Nalan Ǒzkurt; Cuneyt Guzels; Osman Yıldız; Guzels, Cuneyt; Bulucu, Pervin; Yildiz, Osman; Ozkurt, Nalan
This paper presents a convolutional neural network algorithm for the classification of beef samples electronic nose dataset. Proposed algorithm was tested and results were compared to other works that used the same dataset. Overall proposed algorithm showed high performance results without any pre-processing steps. © 2022 Elsevier B.V. All rights reserved.
Citation - WoS: 2
Citation - Scopus: 3
Comparison OF Wavelet Based Feature Extraction Methods for Speech/Music Discrimination
(ISTANBUL UNIV FAC ENGINEERING, 2011) Timur Duzenli; Nalan Ozkurt; Duzenli, Timur; Ozkurt, Nalan
The speech/music discrimination systems have gaining importance in several intelligent audio retrieval algorithms due to the increasing size of the multimedia sources in our daily lives. This study aims to propose a speech/music discrimination system which utilizes the advantages of the wavelet transform. Also the performance of the discrete wavelet transform and the dual-tree wavelet transform has been compared with the conventional time frequency and cepstral domain features used in speech/music discrimination. The speech and music samples collected from common databases CD recording and internet radios have been classified with artificial neural networks with different feature sets. The principal component analysis has been applied to eliminate the correlated features before classification stage. Considering the number of vanishing moments and orthogonality the best performance has been obtained with Daubechies8 wavelet among the other members of the Daubechies family. According to the results the proposed feature set outperforms the traditional ones.
Citation - Scopus: 78
Convolutional Neural Network Hyperparameter Tuning with Adam Optimizer for ECG Classification
(Institute of Electrical and Electronics Engineers Inc., 2020) Sena Yagmur Sen; Nalan Ǒzkurt; Sen, Sena Yagmur; Ozkurt, Nalan
In this research Adaptive Moment Estimation (Adam) optimization technique has been examined on ECG arrhythmia data that rely on deep neural networks. The proposed method indicates that Adam has great importance to solve deep learning problems. According to the proposed method the heartbeats are classified as normal (N) left bundle branch block (LBBB) and right bundle branch block (RBBB) considering the hyper-parameter tuning of the convolutional neural network (CNN). The heartbeats are transformed into spectrogram images and directly given into CNN without any feature extraction method but bounded with a specific frequency/time-resolution rate. The most important point of the study is the examination of the moment estimation coefficients of Adam optimizer such as first moment and second moments. Other tuned parameters are adaptive learning rate and epsilon value. The hyperparameters such as the learning rate and the moment estimation are investigated by grid search method. The effect of the parameters to validation loss were presented and analyzed as a result of this study. © 2020 Elsevier B.V. All rights reserved.
Detection and Semantic Segmentation of Atrial Fibrillation Signals Using U-Net Model
(Institute of Electrical and Electronics Engineers Inc., 2023) Deniz Kan; Selin Gezen; Sevval Nur Canbaz; Nalan Ǒzkurt; Gezen, Selin; Kan, Deniz; Canbaz, Sevval Nur; Ozkurt, Nalan
An abnormal heart rhythm called atrial fibrillation is a dysfunction in the cardiacconduction system that is often life-threatening or reduces the quality of life. This study aims to develop a custom-designed artificial intelligence model UNet to determine whether patients have atrial fibrillation (AF) disease. The purpose is fast and accurate detection. The MIT-BIH Atrial Fibrillation dataset available on the Kaggle platform was used for performance evaluation. The MITBIH Atrial Fibrillation Database includes 25 long-term ECG recordings, however in this project only 18 patient recordings are used due to some of the recordings being unreadable. Two channel ECG signals recorded at 250 samples per second are included in the individual recordings from 18 patients each lasting 10 hours. All simulations were implemented with Phyton in the Kaggle environment. To get the best result the model structure of our U-Net model and parameters such as the activation function and batch size were selected heuristically. Several experiments were done and the model's performance was observed for the different loss functions optimizers and metrics. Our system reaches an accuracy of 93.91% a precision of 99.86% and a recall of 82.40% which results in an F1 score of 90.30%. © 2023 Elsevier B.V. All rights reserved.
Detection of Atrial Fibrillation with Custom Designed Wavelet-based Convolutional Autoencoder
(2024) Evrim Şimşek; NALAN OZKURT; Öykü Eravcı; Özlem Memiş; Şimşek, Evrim; Eravcı, Öykü; Memiş, Özlem; Ozkurt, Nalan
Remote monitoring of patients is of great importance in terms of early diagnosis of diseases and improving people's quality of life. With the rapid development of deep learning techniques wearable health technologies have leaped forward. This has made the automatic diagnosis even more important. In this study we provide a deep learning approach for classifying Atrial Fibrillation (AF) arrhythmia that uses a customized wavelet-based convolutional autoencoder (WCAE) model. WCAE is employed as an anomaly detector which combines the time-frequency domain examination ability of wavelet and the data-driven feature learning capability of convolutional autoencoders. The proposed approach received average scores of 95.45% 99.99% 90.90% and 95.23% for accuracy precision recall and F1 respectively on a large selection of publicly available datasets. The outcomes of the experiments demonstrate the significance of using deep learning-based models in diagnosing AF. Moreover it is observed that utilization of wavelet methods along with autoencoder model has a great potential for biomedical signal processing systems.
Citation - WoS: 11
Citation - Scopus: 27
ECG Arrhythmia Classification By Using Convolutional Neural Network And Spectrogram
(IEEE, 2019) Sena Yagmur Sen; Nalan Ozkurt; Sen, Sena Yagmur; Ozkurt, Nalan
In this study the electrocardiography (ECG) arrhythmias have been classified by the proposed framework depend on deep neural networks in order to features information. The proposed approaches operates with a large volume of raw ECG time-series data and ECG signal spectrograms as inputs to a deep convolutional neural networks (CNN). Heartbeats are classified as normal ( N) premature ventricular contractions (PVC) right bundle branch block (RBBB) rhythm by using ECG signals obtained from MIT-BIH arrhythmia database. The first approach is to directly use ECG time-series signals as input to CNN and in the second approach ECG signals are converted into time-frequency domain matrices and sent to CNN. The most appropriate parameters such as number of the layers size and number of the filters are optimized heuristically for fast and efficient operation of the CNN algorithm. The proposed system demonstrated high classification rate for the time-series data and spectrograms by using deep learning algorithms without standard feature extraction methods. Performance evaluation is based on the average sensitivity specificity and accuracy values. It is also worth to note that spectrogram increases the performance of classification since it extracts the useful time-frequency information of the signal.
Citation - Scopus: 1
ECG Arrhythmia Detector with Custom Designed Wavelet-Based Convolutional Autoencoder for Unbalanced Data
(Institute of Electrical and Electronics Engineers Inc., 2025) Eravci, Oyku; Sarvan, Cagla; Ozkurt, Nalan
Citation - WoS: 4
Citation - Scopus: 14
ECG beat arrhythmia classification by using 1-d CNN in case of class imbalance
(Institute of Electrical and Electronics Engineers Inc., 2019) Çağla Sarvan; Nalan Ǒzkurt; Sarvan, Cagla; Ozkurt, Nalan
In this study ECG arrhythmia types of non-ectopic (N) ventricular ectopic (V) unknown (Q) supraventricular ectopic (S) and fusion (F) were classified by using the convolutional neural network (CNN) architecture. QRS detection was performed on these ECG arrhythmias that downloaded from MIT-BIH database. An imbalanced number of beats was obtained for 5 different arrhythmia types. In order to reduce the effect of imbalance in statistical performance metrics data mining techniques such as recall of data were applied. It was aimed to increase the positive predictive value (PPV) rates of the classes which consist of a few instances. © 2020 Elsevier B.V. All rights reserved.
Citation - WoS: 7
Citation - Scopus: 11
EEG motor movement classification based on cross-correlation with effective channel
(SPRINGER LONDON LTD, 2019) Mohand Lokman Al-dabag; Nalan Ozkurt; Al-dabag, Mohand Lokman; Ozkurt, Nalan
In brain-computer interface (BCI) systems the classification of electroencephalography (EEG) mental tasks is an important issue. This classification involves many steps: signal preprocessing feature extraction and classification. In this study a simple and robust method is proposed for preprocessing and feature extraction stages of the EEG classification. The method includes noise removal by EEG subtraction channel selection EEG band extraction using discrete wavelet transform cross-correlation of EEG channels with effective channels and statistical parameter calculation. Two datasets are classified to illustrate the performance of the proposed method. One of them is the BCI competition III dataset IVa which is commonly used in research articles and the second is recorded using Emotiv Epoc+headset. The results show that the average accuracy of the classification using an artificial neural network and support vector machine is above 96%.
Citation - WoS: 21
Citation - Scopus: 48
Emotion Classification from EEG Signals in Convolutional Neural Networks
(IEEE, 2019) Hayriye Donmez; Nalan Ozkurt; Donmez, Hayriye; Ozkurt, Nalan
The objective of this research is to classify EEG (electroencephalography) signal recordings of the subjects evoked by visual stimulus by using CNN (Convolutional Neural Networks). EEG records the electrical activity of brain signals. In medicine EEG is used to diagnose some neurological disorders but moreover the classification of the emotions is also possible from EEG recordings. Emotion recognition is an important task for the computers in machine perception. Therefore in this study the participants are presented with a video containing funny scary and sad excerpts and simultaneously EEG signal is measured by Neurosky Mindwave EEG Headset. The spectrogram of EEG signals is supplied to CNN and three emotions are classified using brain signal spectrogram images.
Enhancing ADHD Detection via Functional Connectivity: Autoencoder-Based Feature Selection and DMN ROI Focus
(Springer Science and Business Media Deutschland GmbH, 2025) Taspinar, Gurcan; Ozkurt, Nalan
Citation - WoS: 4
Citation - Scopus: 10
Feature Selection and Classification of EEG Finger Movement Based on Genetic Algorithm
(IEEE, 2018) Mohand Lokman Al Dabag; Nalan Ozkurt; Shaima Miqdad Mohamed Najeeb; Najeeb, Shaima Miqdad Mohamed; Ozkurt, Nalan; Al Dabag, Mohand Lokman; BM Ozyildirim; T Yildirim
Electroencephalography (EEG) classification for mental tasks is the crucial part of the brain-computer interface. Many studies try to extract discriminative features from EEG signals. In this study feature selection algorithm based on genetic algorithm (GA) was implemented to find the best features that describe EEG signal. The best features are searched among ten statistical features calculated from the cross-correlation of effective channel with relevant EEG channels in the proposed study. A comparison was made after and before feature selection in two major viewpoints: classification accuracy and computation time. Multi-Layer Perceptron Neural Network (MLP) and Support Vector Machine (SVM) are used to classify left and right finger movements of 13 subjects. The overall classification performance is enhanced about 1% for both classifiers after feature selection. Computation time has reduced about 34% in SVM classifier and there is huge reduction about 84% in MLP.
Citation - WoS: 2
Citation - Scopus: 2
Feature Selection for ECG Beat Classification using Genetic Algorithms with A Multi-objective Approach
(IEEE, 2018) Cagla Sarvan; Nalan Ozkurt; Sarvan, Cagla; Ozkurt, Nalan
To identify appropriate features in classification studies is a common problem in many areas. In this study a genetic algorithm method with multi-objective approach is proposed for selecting the features that give high performance ratio in classifying cardiac arrhythmia. Discrete Wavelet Transform (DWT) were used for extracting features from Normal right bundle branch block left bundle branch block and paced rhythm recordings of electrocardiography (ECG) signals which were taken from the MIT-BIH cardiac arrhythmia database. Using 13 different wavelet types 208 features were obtained by the DWT method. Among these features a minimum number of feature sets were chosen to provide high performance in classification. Then the classification results were compared with the results of the classical genetic algorithm which aims to improve accuracy.
FMRG Hacimlerini Kullanarak DEHB’nin 3B ESA Tabanlı Otomatik Teşhisi
(2023) NALAN OZKURT; Gurcan Taspinar; Taspinar, Gurcan; Ozkurt, Nalan
Dikkat eksikliği hiperaktivite bozukluğu (DEHB) en sık görülen beyin deformasyonlarından biridir ve özellikle çocukların okul başarılarını olumsuz yönde etkilemektedir. Uzmanların DEHB hastalarına uygun tedavi verebilmeleri için bu hastalığın nörobiyolojik tanısı önemlidir. Makine öğrenimi algoritmaları ile birlikte nörogörüntüleme teknolojileri özellikle fonksiyonel manyetik rezonans görüntüleme dikkat eksikliği hiperaktivite bozukluğunda biyobelirteç olarak giderek daha fazla kullanılmaktadır. Ayrıca makine öğrenme yöntemleri son zamanlarda popüler hale gelmektedir. Bu çalışma ile DEHB tanısında uzmanlara yardımcı olmak amacıyla fonksiyonel manyetik rezonans görüntüleme hacimlerini iki sınıfa ayırmak için optimize edilmiş 3 boyutlu evrişimli bir sinir ağı sunulmaktadır. Verilerin 3 boyutlu ilişkilerinin çıkarılmasının önemini göstermek için yöntem halka açık ADHD-200 veri setlerinin öğrenme ve test verileri kullanılarak test edilmiş ve sinir ağının performansı değerlendirilmiştir. Daha sonra sinir ağının performansı literatürdeki birkaç yeni DEHB algılama evrişimli sinir ağı ile karşılaştırılmıştır. Kullanılan sinir ağının umut verici bir performansa sahip olduğu gözlemlenmektedir.