Bilgisayar Mühendisliği Bölümü

Permanent URI for this communityhttps://hdl.handle.net/20.500.12416/13

Browse

Now showing 1 - 20 of 42

Citation - WoS: 32
Citation - Scopus: 43
A 3d Virtual Environment for Training Soccer Referees
(Elsevier Science Bv, 2019) Isler, Veysi; O'Connor, Rory V.; Clarke, Paul M.; Gulec, Ulas; Yilmaz, Murat
Emerging digital technologies are being used in many ways by and in particular virtual environments provide new opportunities to gain experience on real-world phenomena without having to live the actual real-world experiences. In this study, a quantitative research approach supported by expert validation interviews was conducted to determine the availability of virtual environments in the training of soccer referees. The aim is to design a virtual environment for training purposes, representing a real-life soccer stadium to allow the referees to manage matches in an atmosphere similar to the real stadium atmosphere. At this point, the referees have a chance to reduce the number of errors that they make in real life by experiencing difficult decisions that they encounter during the actual match via using the virtual stadium. In addition, the decisions and reactions of the referees during the virtual match were observed with the number of different fans in the virtual stadium to understand whether the virtual stadium created a real stadium atmosphere for the referees. For this evaluation, Presence Questionnaire (PQ) and Immersive Tendencies Questionnaire (ITQ) were applied to the referees to measure their involvement levels. In addition, a semi-structure interview technique was utilized in order to understand participants' opinions about the system. These interviews show that the referees have a positive attitude towards the system since they can experience the events occurred in the match as a first person instead of watching them from camera as a third person. The findings of current study suggest that virtual environments can be used as a training tool to increase the experience levels of the soccer referees since they have an opportunity to decide about the positions without facing the real-world risks.
Citation - WoS: 4
Citation - Scopus: 5
Almost Autonomous Training of Mixtures of Principal Component Analyzers
(Elsevier Science Bv, 2004) Musa, MEM; de Ridder, D; Duin, RPW; Atalay, V
In recent years, a number of mixtures of local PCA models have been proposed. Most of these models require the user to set the number of submodels (local models) in the mixture and the dimensionality of the submodels (i.e., number of PC's) as well. To make the model free of these parameters, we propose a greedy expectation-maximization algorithm to find a suboptimal number of submodels. For a given retained variance ratio, the proposed algorithm estimates for each submodel the dimensionality that retains this given variability ratio. We test the proposed method on two different classification problems: handwritten digit recognition and 2-class ionosphere data classification. The results show that the proposed method has a good performance. (C) 2004 Elsevier B.V. All rights reserved.
Analysing Iraqi Railways Network by Applying Specific Criteria Using the Gis Techniques
(Coll Science Women, Univ Baghdad, 2019) Naji, Hayder Fans; Maras, H. Hakan
The railways network is one of the huge infrastructure projects. Therefore, dealing with these projects such as analyzing and developing should be done using appropriate tools, i.e. GIS tools. Because, traditional methods will consume resources, time, money and the results maybe not accurate. In this research, the train stations in all of Iraq's provinces were studied and analyzed using network analysis, which is one of the most powerful techniques within GIS. A free trial copy of ArcGIS (R) 10.2 software was used in this research in order to achieve the aim of this study. The analysis of current train stations has been done depending on the road network, because people used roads to reach those train stations. The data layers for this study were collected and prepared to meet the requirements of network analyses within GIS. In this study, the current train stations in Iraq were analyzed and studied depending on accessibility value for those stations. Also, to know the numbers of people who can reach those stations within a walking time of 20 minutes. So, this study aims to analyze the current train stations according to multiple criteria by using network analysis in order to find the serviced areas around those stations. Results will be presented as digital maps layers with their attribute tables that show the beneficiaries from those train stations and serviced areas around those stations depending on specific criteria, with a view to determine the size of this problem and to support the decision makers in case of locating new train stations within the best locations for it.
Citation - WoS: 16
Citation - Scopus: 20
Application of Bilstm-Crf Model With Different Embeddings for Product Name Extraction in Unstructured Turkish Text
(Springer London Ltd, 2024) Arslan, Serdar
Named entity recognition (NER) plays a pivotal role in Natural Language Processing by identifying and classifying entities within textual data. While NER methodologies have seen significant advancements, driven by pretrained word embeddings and deep neural networks, the majority of these studies have focused on text with well-defined grammar and structure. A significant research gap exists concerning NER in informal or unstructured text, where traditional grammar rules and sentence structure are absent. This research addresses this crucial gap by focusing on the detection of product names within unstructured Turkish text. To accomplish this, we propose a deep learning-based NER model which combines a Bidirectional Long Short-Term Memory (BiLSTM) architecture with a Conditional Random Field (CRF) layer, further enhanced by FastText embeddings. To comprehensively evaluate and compare our model's performance, we explore different embedding approaches, including Word2Vec and Glove, in conjunction with the Bidirectional Long Short-Term Memory and Conditional Random Field (BiLSTM-CRF) model. Furthermore, we conduct comparisons against BERT to assess the efficacy of our approach. Our experimentation utilizes a Turkish e-commerce dataset gathered from the internet, where traditional grammatical and structural rules may not apply. The BiLSTM-CRF model with FastText embeddings achieved an F1 score value of 57.40%, a precision value of 55.78%, and a recall value of 59.12%. These results indicate promising performance in outperforming other baseline techniques. This research contributes to the field of NER by addressing the unique challenges posed by unstructured Turkish text and opens avenues for improved entity recognition in informal language settings, with potential applications across various domains.
Citation - WoS: 3
Citation - Scopus: 3
Binary Background Model With Geometric Mean for Author-Independent Authorship Verification
(Sage Publications Ltd, 2023) Sezer, Ebru A.; Sever, Hayri; Canbay, Pelin
Authorship verification (AV) is one of the main problems of authorship analysis and digital text forensics. The classical AV problem is to decide whether or not a particular author wrote the document in question. However, if there is one and relatively short document as the author's known document, the verification problem becomes more difficult than the classical AV and needs a generalised solution. Regarding to decide AV of the given two unlabeled documents (2D-AV), we proposed a system that provides an author-independent solution with the help of a Binary Background Model (BBM). The BBM is a supervised model that provides an informative background to distinguish document pairs written by the same or different authors. To evaluate the document pairs in one representation, we also proposed a new, simple and efficient document combination method based on the geometric mean of the stylometric features. We tested the performance of the proposed system for both author-dependent and author-independent AV cases. In addition, we introduced a new, well-defined, manually labelled Turkish blog corpus to be used in subsequent studies about authorship analysis. Using a publicly available English blog corpus for generating the BBM, the proposed system demonstrated an accuracy of over 90% from both trained and unseen authors' test sets. Furthermore, the proposed combination method and the system using the BBM with the English blog corpus were also evaluated with other genres, which were used in the international PAN AV competitions, and achieved promising results.
Citation - WoS: 23
Citation - Scopus: 32
A Compact Multiband Printed Monopole Antenna With Hybrid Polarization Radiation for Gps, Lte, and Satellite Applications
(Ieee-inst Electrical Electronics Engineers inc, 2020) Al-Mihrab, Mohammed A.; Salim, Ali J.; Ali, Jawad K.
A new compact printed monopole antenna is presented in this paper. An open-loop hexagonal radiator excited by a microstrip feed line, which is printed on top of the substrate, which is FR4 type, while on another side, a partial ground plane is fixed and embedded with two pairs of slits as well as a pair of rectangular strips. Triple operating bands with two different polarization types are obtained. The lower band has right-hand circular polarization (RHCP) characteristic, whereas the upper band has left-hand circular polarization (LHCP) characteristic means that a dual-band dual-sense circular polarization (CP). Concerning the middle band, a linear polarization (LP) has been gotten in this antenna. Numerical analysis and experimental validation of the proposed antenna structure have been performed, and results are demonstrated. The measured impedance bandwidths (IBWs) are 14.7% (1.478-1.714 GHz), 6.8% (2.54-2.72 GHz), and 13.1% (4.29-4.89 GHz), respectively. The measured 3-dB axial ratio bandwidths (ARBWs) are 6.2% (1.510-1.606 GHz), and 22.7% (4.035-5.07 GHz) for the lower and the upper band, respectively. So, it's suitable for covering modern wireless applications such as GPS (Global Positioning System), LTE (Long Term Evaluation), and Satellite.
Citation - WoS: 8
Citation - Scopus: 8
Comparative Analysis on Wavelet-Based Detection of Finite Duration Low-Amplitude Signals Related To Ventricular Late Potentials
(Iop Publishing Ltd, 2004) Mousa, A; Yilmaz, A
Ventricular late potentials (VLPs) are considered as a noninvasive marker of patients with myocardial infarction, who are prone to the development of ventricular tachycardia. This paper investigates the effects of variations in physical properties of myocardial infarcts in terms of their effects on the parametric variations in VLP analysis. A sufficiently large set of signals underlining the behavior of physical parameters was employed to represent the effect of physical size, position, orientation and type of infarct. The approximated signals are variations from real electrocardiography signals by adding potentials representing late potentials based on duration, frequency, amplitude and position. The aim is not to exactly model VLP but rather to generate an approximate set of signals to examine the performance of the standard methods for different possibilities in infarct dynamics. We investigate some of the detection approaches together with their related assumptions, and try to pinpoint the drawbacks and inaccuracies of these methods and also their assumptions. The three widely accepted criteria-QRS duration, root-mean-square and duration of the signal at the end of QRS for VLP detection-were used in the investigation. Results from the application of these parameters to the set of signals are presented. In addition we investigate the physical nature of an infarct and list a number of possible reasons that might be the cause of a low success rate for the detection of additive potentials. To improve the performance of the common methods, two more wavelet transform parameters are added to those of the standard methods. The method derived from this analysis is presented as an alternative means for the detection of late signals named as delayed potentials, a more general class that includes VLP as a subset.
D2 +Nin(T), n=7 and 9, Collision System
(1999) Böyükata, Mustafa; Durmuş, Perihan; Özçelik, Süleyman; Güvenç, Ziya Burhanettin; Jellinek, Julius
In this study the kinetics of reactions of Nin n=7 and 9, clusters with a deuterium D2 molecule are studied via quasiclassical molecular dynamics. Dissociative chemisorption probabilities as functions of impact parameters, collision energies, and a rovibrational state of a molecule are calculated. And the corresponding reactive cross sections are evaluated. Resonance formation in the low collision energy region is discussed.
Citation - WoS: 42
Citation - Scopus: 42
A Density Functional Study of Bare and Hydrogenated Platinum Clusters
(Elsevier, 2006) Sebetci, Ali
We perform density functional theory calculations using Gaussian atomic-orbital methods within the generalized gradient approximation for the exchange and correlation to study the interactions in the bare and hydrogenated platinum clusters. The minimum-energy structures, binding energies, relative stabilities. vibrational frequencies and the highest occupied and lowest unoccupied molecular-orbital gaps of PtnHm (n = 1-5, m = 0-2) clusters are calculated and compared with previously studied pure platinum and hydrogenated platinum clusters. We investigate any magic behavior in hydrogenated platinum clusters and find that Pt4H2 is snore stable than its neighboring sizes. The lowest energy structure of Pt-4 is found to be a distorted tetrahedron and that of Pt-5 found to be a bridge site capped tetrahedron which is a new global minimum for Pt-5 cluster. The successive addition of H atoms to Pt-n clusters leads to an oscillatory change in the magnetic moment of Pt-3-Pt-5 clusters. (c) 2006 Elsevier B.V. All rights reserved.
Citation - WoS: 42
Citation - Scopus: 49
Detection of Hip Osteoarthritis by Using Plain Pelvic Radiographs With Deep Learning Methods
(Springer, 2020) Ureten, Kemal; Arslan, Tayfun; Gultekin, Korcan Emre; Demir, Ayse Nur Demirgoz; Ozer, Hafsa Feyza; Bilgili, Yasemin
Objective The incidence of osteoarthritis is gradually increasing in public due to aging and increase in obesity. Various imaging methods are used in the diagnosis of hip osteoarthritis, and plain pelvic radiography is the first preferred imaging method in the diagnosis of hip osteoarthritis. In this study, we aimed to develop a computer-aided diagnosis method that will help physicians for the diagnosis of hip osteoarthritis by interpreting plain pelvic radiographs. Materials and methods In this retrospective study, convolutional neural networks were used and transfer learning was applied with the pre-trained VGG-16 network. Our dataset consisted of 221 normal hip radiographs and 213 hip radiographs with osteoarthritis. In this study, the training of the network was performed using a total of 426 hip osteoarthritis images and a total of 442 normal pelvic images obtained by flipping the raw data set. Results Training results were evaluated with performance metrics such as accuracy, sensitivity, specificity, and precision calculated by using the confusion matrix. We achieved accuracy, sensitivity, specificity and precision results at 90.2%, 97.6%, 83.0%, and 84.7% respectively. Conclusion We achieved promising results with this computer-aided diagnosis method that we tried to develop using convolutional neural networks based on transfer learning. This method can help clinicians for the diagnosis of hip osteoarthritis while interpreting plain pelvic radiographs, also provides assistance for a second objective interpretation. It may also reduce the need for advanced imaging methods in the diagnosis of hip osteoarthritis.
Citation - WoS: 62
Citation - Scopus: 70
Detection of Rheumatoid Arthritis From Hand Radiographs Using a Convolutional Neural Network
(Springer London Ltd, 2020) Ureten, Kemal; Erbay, Hasan; Maras, Hadi Hakan
Introduction Plain hand radiographs are the first-line and most commonly used imaging methods for diagnosis or differential diagnosis of rheumatoid arthritis (RA) and for monitoring disease activity. In this study, we used plain hand radiographs and tried to develop an automated diagnostic method using the convolutional neural networks to help physicians while diagnosing rheumatoid arthritis. Methods A convolutional neural network (CNN) is a deep learning method based on a multilayer neural network structure. The network was trained on a dataset containing 135 radiographs of the right hands, of which 61 were normal and 74 RA, and tested it on 45 radiographs, of which 20 were normal and 25 RA. Results The accuracy of the network was 73.33% and the error rate 0.0167. The sensitivity of the network was 0.6818; the specificity was 0.7826 and the precision 0.7500. Conclusion Using only pixel information on hand radiographs, a multi-layer CNN architecture with online data augmentation was designed. The performance metrics such as accuracy, error rate, sensitivity, specificity, and precision state shows that the network is promising in diagnosing rheumatoid arthritis.
Citation - WoS: 8
Citation - Scopus: 11
The Diagnosis of Femoroacetabular Impingement Can Be Made on Pelvis Radiographs Using Deep Learning Methods
(Turkish Joint Diseases Foundation, 2023) Atalar, Ebru; Ureten, Kemal; Kanatli, Ulunay; Ciceklidag, Murat; Kaya, Ibrahim; Vural, Abdurrahman; Maras, Yuksel
Objectives: The aim of this study was to evaluate diagnostic ability of deep learning models, particularly convolutional neural network models used for image classification, for femoroacetabular impingement (FAI) using hip radiographs. Materials and methods: Between January 2010 and December 2020, pelvic radiographs of a total of 516 patients (270 males, 246 females; mean age: 39.1 +/- 3.8 years; range, 20 to 78 years) with hip pain were retrospectively analyzed. Based on inclusion and exclusion criteria, a total of 888 hip radiographs (308 diagnosed with FAI and 508 considered normal) were evaluated using deep learning methods. Pre-trained VGG-16, ResNet-101, MobileNetV2, and Inceptionv3 models were used for transfer learning. Results: As assessed by performance measures such as accuracy, sensitivity, specificity, precision, F-1 score, and area under the curve (AUC), the VGG-16 model outperformed other pre-trained networks in diagnosing FAI. With the pre-trained VGG-16 model, the results showed 86.6% accuracy, 82.5% sensitivity, 89.6% specificity, 85.5% precision, 83.9% F1 score, and 0.92 AUC. Conclusion: In patients with suspected FAI, pelvic radiography is the first imaging method to be applied, and deep learning methods can help in the diagnosis of this syndrome.
Citation - WoS: 11
Citation - Scopus: 11
Diagnosis of Osteoarthritic Changes, Loss of Cervical Lordosis, and Disc Space Narrowing on Cervical Radiographs With Deep Learning Methods
(Turkish Joint Diseases Foundation, 2022) Tokdemir, Gul; Ureten, Kemal; Atalar, Ebru; Duran, Semra; Maras, Hakan; Maras, Yuksel
Objectives: In this study, we aimed to differentiate normal cervical graphs and graphs of diseases that cause mechanical neck pain by using deep convolutional neural networks (DCNN) technology. Materials and methods: In this retrospective study, the convolutional neural networks were used and transfer learning method was applied with the pre-trained VGG-16, VGG-19, Resnet-101, and DenseNet-201 networks. Our data set consisted of 161 normal lateral cervical radiographs and 170 lateral cervical radiographs with osteoarthritis and cervical degenerative disc disease. Results: We compared the performances of the classification models in terms of performance metrics such as accuracy,
Citation - WoS: 8
Citation - Scopus: 9
Effects of Hydrogen Hosting on Cage Structures of Boron Clusters: Density Functional Study of Bmhn (m=5-10 and N ≤ M) Complexes
(Iop Publishing Ltd, 2008) Ozdogan, C.; Guvenc, Z. B.; Boyukata, M.
The structural stability of hydrogen bonded boron microclusters has been studied by using the density functional theory. Effects of the increasing number of hydrogen atoms on the cage geometries of B-5-B-10 clusters, and the distortion of the cage configurations of the boranes are assessed. The possible stable structures of BmHn(m = 5-10 and n <= m) boron hydrides, their binding energies, HOMO-LUMO energy gaps and the total atomic charges of the B-m in the complexes are determined. For the series of B5Hn, B7Hn, and B9Hn major structural changes are observed.
Citation - WoS: 3
Citation - Scopus: 3
Finite Size Scaling by Using Scaling Functions in Two-Dimensional Q=2 and 7 State Potts Models
(World Scientific Publ Co Pte Ltd, 2001) Seferoglu, N; Aydin, M; Gündüç, Y; Demirtürk, S
The scaling behaviors of the percolation cumulant and the surface renormalization are studied on q = 2 and 7 state Potts models. The results show that the scaling functions can be safely used to determine infinite lattice transition points and the thermal and magnetic exponents indicating that these functions have very small correction to scaling contributions.
Fragmentation of a Non-Rotating Ni19 Cluster: A Molecular Dynamics Study
(1999) Avcı, Halil; Çivi, Mehmet; Güvenç, Ziya Burhanettin; Jellinek, Julius
Collisionless fragmentation of a non-rotating Ni19 cluster is studied using constant-energy molecular dynamics computer simulations. The cluster is modelled by an embedded atom model (EAM) energy surface. Distribution of the channel-specific fragmentation probabilities, and global rate constants are computed and analyzed as functions of the internal energy of the cluster. The results are compared with those obtained using the RRK statistical approach, and also compared with the other multi-channel fragmentation work.
Citation - WoS: 22
Citation - Scopus: 21
Functionalizing Graphene by Embedded Boron Clusters
(Iop Publishing Ltd, 2008) Ozdogan, Cem; Kunstmann, Jens; Fehske, Holger; Quandt, Alexander
We present a model system that might serve as a blueprint for the controlled layout of graphene based nanodevices. The systems consists of chains of B-7 clusters implanted in a graphene matrix, where the boron clusters are not directly connected. We show that the graphene matrix easily accepts these alternating B-7-C-6 chains and that the implanted boron components may dramatically modify the electronic properties of graphene based nanomaterials. This suggests a functionalization of graphene nanomaterials, where the semiconducting properties might be supplemented by parts of the graphene matrix itself, but the basic wiring will be provided by alternating chains of implanted boron clusters that connect these areas.
Low-diameter topic-based pub/sub overlay Network Construction with minimum–maximum node Degree
(2021) Yumusak, Semih; Layazali, Sina; Öztoprak, Kasım; Hassanpour, Reza
In the construction of effective and scalable overlay networks, publish/subscribe (pub/sub) network designers prefer to keep the diameter and maximum node degree of the network low. However, existing algorithms are not capable of simultaneously decreasing the maximum node degree and the network diameter. To address this issue in an overlay network with various topics, we present herein a heuristic algorithm, called the constant-diameter minimum–maximum degree (CD-MAX), which decreases the maximum node degree and maintains the diameter of the overlay network at two as the highest. The proposed algorithm based on the greedy merge algorithm selects the node with the minimum number of neighbors. The output of the CD-MAX algorithm is enhanced by applying a refinement stage through the CD-MAX-Ref algorithm, which further improves the maximum node degrees. The numerical results of the algorithm simulation indicate that the CD-MAX and CD-MAX-Ref algorithms improve the maximum node-degree by up to 64% and run up to four times faster than similar algorithms.
Citation - WoS: 13
Citation - Scopus: 17
A Mash-Up Application Utilizing Hybridized Filtering Techniques for Recommending Events at a Social Networking Site
(Springer Wien, 2011) Kayaalp, Mehmet; Ozyer, Tansel; Ozyer, Sibel T.
Event recommendation is one way of gathering people having same likes/dislikes. In today's world, many mass amounts of events are organized at different locations and times. Generally, cliques of people are fans of some specific events. They attend together based on each other's recommendation. Generally, there are many activities that people prefer/opt out attending and these events are announced for attracting relevant people. Rather than, peerto-peer oracles of a local group of people, or sentiments of people from different sources, an intelligent recommendation system can be used at a social networking site in order to recommend people in collaborative and content basis within a social networking site. We have used an existing social environment (http://www.facebook.com) for deployment. Our application has also been integrated with several web sites for collecting information for assessment. Our system has been designed in modules so that it is open to new data sources either by using web services or web scraping. Currently, our application is yet an application that permits users rate events; they have attended or have beliefs on them. Given the social network between people, system tries to recommend upcoming events to users. For this purpose, we have exploited the fact that a similarity relationship between different events can exist in terms of both content and collaborative filtering. Geographical locations have an impact so; we have also taken geographical location information and social concept of an event. Eventually, our system integrates different sources in facebook (http://www.facebook.com) for doing recommendation between people in close relationship. We have performed experiments among a group of students. Experiments led us have promising results.
Citation - WoS: 58
Citation - Scopus: 66
Molecular Dynamics Simulation of Sintering and Surface Premelting of Silver Nanoparticles
(Japan inst Metals & Materials, 2013) Ozdogan, C.; Hu, A.; Yavuz, M.; Zhou, Y.; Atis, M.; Alarifi, H. A.
Sintering of Ag nanoparticles (NPs) is increasingly being used as a driving mechanism for joining in the microelectronics industry. We therefore performed molecular dynamics simulations based on the embedded atom method (EAM) to study pressureless sintering kinetics of two Ag NPs in the size range of (4 to 20 nm), and sintering of three and four Ag NPs of 4 nm diameter. We found that the sintering process passed through three main stages. The first was the neck formation followed by a rapid increase of the neck radius at 50K for 20 nm particles and at 10 K for smaller NPs. The second was characterized by a gradual linear increase of the neck radius to particle radius ratio as the temperature of the sintered structure was increased to the surface premelting point. Different than previous sintering studies, a twin boundary was formed during the second stage that relaxed the sintered structure and decreased the average potential energy (PE). The third stage of sintering was a rapid shrinkage during surface premelting of the sintered structure. Based on pore geometry, densification occurred during the first stage for three 4 nm particles and during the second stage for four 4 nm particles. Sintering rates obtained by our simulation were higher than those obtained by theoretical models generally used for predicting sintering rates of microparticles.

Browse

Browsing Bilgisayar Mühendisliği Bölümü by WoS Q "Q2"