Yazılım Mühendisliği Bölümü

Permanent URI for this communityhttps://hdl.handle.net/20.500.12416/2146

Browse

Now showing 1 - 20 of 38

Otomatik Konuşma Tanımaya Genel Bakış, Yaklaşımlar ve Zorluklar: Türkçe Konuşma Tanımanın Gelecekteki Yolu
(2019) Oyucu, Saadin; Polat, Huseyin; Sever, Hayri
İnsanlar arasındaki en önemli iletişim yöntemi olan konuşmanın, bilgisayarlar tarafından tanınması önemli bir çalışma alanıdır. Bu araştırma alanında farklı diller temel alınarak birçok çalışma gerçekleştirilmiştir. Literatürdeki çalışmalar konuşma tanıma teknolojilerinin başarımının artmasında önemli rol oynamıştır. Bu çalışmada konuşma tanıma ile ilgili bir literatür taraması yapılmış ve detaylı olarak sunulmuştur. Ayrıca farklı dillerde bu araştırma alanında kaydedilen ilerlemeler tartışılmıştır. Konuşma tanıma sistemlerinde kullanılan veri setleri, özellik çıkarma yaklaşımları, konuşma tanıma yöntemleri ve performans değerlendirme ölçütleri incelenerek konuşma tanımanın gelişimi ve bu alandaki zorluklara odaklanılmıştır. Konuşma tanıma alanında son zamanlarda yapılan çalışmaların olumsuz koşullara (çevre gürültüsü, konuşmacıda ve dilde değişkenlik) karşı çok daha güçlü yöntemler geliştirmeye odaklandığı izlenmiştir. Bu nedenle araştırma alanı olarak genişleyen olumsuz koşullardaki konuşma tanıma ile ilgili yakın geçmişteki gelişmelere yönelik genel bir bakış açısı sunulmuştur. Böylelikle olumsuz koşullar altında gerçekleştirilen konuşma tanımadaki tıkanıklık ve zorlukları aşabilmek için kullanılabilecek yöntemleri seçmede yardımcı olunması amaçlanmıştır. Ayrıca Türkçe konuşma tanımada kullanılan ve iyi bilinen yöntemler karşılaştırılmıştır. Türkçe konuşma tanımanın zorluğu ve bu zorlukların üstesinden gelebilmek için kullanılabilecek uygun yöntemler irdelenmiştir. Buna bağlı olarak Türkçe konuşma tanımanın gelecekteki rotasına ilişkin bir değerlendirme ortaya konulmuştur.
Tacıt Knowledge Vısualızatıon Through Organızatıonal Explıcıt Knowledge Warehouses: A Proposal For Research Methodology Desıgn And Executıon
(2011) Medeni, İhsan Tolga; Medeni, Tunç; Tolun, Mehmet
Knowledge visualization can be used in several fields from medical imaging to industrial engineering. Although there could be variety of applicable research areas, our consideration will be the tacit knowledge visualization in organizations. This proposal aims to suggest a study to develop a tacit knowledge visualization framework to support know-where requirements of the organizational knowledge. With the implementation of our framework in a software application, it is aimed to create a virtual environment, where subject-based knowledge requirements will be answered by the visualized tacit knowledge of individuals and possibly the relations among individual members of the organization
Citation - WoS: 20
Citation - Scopus: 29
Creating Consensus Group Using Online Learning Based Reputation in Blockchain Networks
(Elsevier, 2019) Ozsoy, Adnan; Oztaner, Serdar Murat; Sever, Hayri; Bugday, Ahmet
One of the biggest challenges to blockchain technology is the scalability problem. The choice of consensus algorithm is critical to the practical solution of the scalability problem. To increase scalability, Byzantine Fault Tolerance (BFT) based methods have been most widely applied. This study proposes a new model instead of Proof of Work (PoW) for forming the consensus group that allows the use of BFT based methods in the public blockchain network. The proposed model uses the adaptive hedge method, which is a decision-theoretic online learning algorithm (Qi et al., 2016). The reputation value is calculated for the nodes that want to participate in the consensus committee, and nodes with high reputation values are selected for the consensus committee to reduce the chances of the nodes in the consensus committee being harmful. Since the study focuses on the formation of the consensus group, a simulated blockchain network is used to test the proposed model more effectively. Test results indicate that the proposed model, which is a new approach in the literature making use of machine learning for the construction of consensus committee, successfully selects the node with the higher reputation for the consensus group. (C) 2019 Elsevier B.V. All rights reserved.
Citation - WoS: 20
Citation - Scopus: 23
Investigation of Equatorial Plasma Bubble Irregularities Under Different Geomagnetic Conditions During the Equinoxes and the Occurrence of Plasma Bubble Suppression
(Pergamon-elsevier Science Ltd, 2020) Timocin, Erdinc; Inyurt, Samed; Temucin, Huseyin; Ansari, Kutubuddin; Jamjareegulgarn, Punyawi
In this study, we investigated the behavior of equatorial plasma bubble irregularities under different geomagnetic conditions during March 2015 and September 2017. It was used Total Electron Content (TEC) data obtained from SGOC (6,89 degrees N, 79,87 degrees E), IISC (12,94 degrees N, 77,57 degrees E) and HYDE (17,40 degrees N, 78,50 degrees E) receiver stations which located between the trough and the crest of the equatorial ionization anomaly (EIA). We used the Rate of TEC change (ROT) and Rate of TEC change index (ROTI) to represent plasma bubbles irregularities. These indices are a well proxy for the ionospheric fluctuations and can be used to describe features of plasma bubbles irregularities. The equatorial plasma bubble irregularities for all stations were observed between 13 UT and 20 UT (during postsunset period) during equinoxes. The intensity level of ROTI during postsunset periods was greater than 1 TECU min(-1). Also, the values of mean ROTI (ROTIave) between 13 UT and 20 UT have values greater than 0,4 TECU min(-1) while the values of ROTIave at the other hours have values less than 0,4 TECU min(-1). The geomagnetic activity has a significant effect on the occurrence of equatorial plasma bubbles irregularities. The occurrence rate of equatorial plasma bubble irregularities observed during postsunset hours increased as geomagnetic activity increases. It also was observed that the main phases of geomagnetic storms have the triggering effect of storms on equatorial plasma bubble irregularities observed at postsunset hours while the recovery phases of geomagnetic storms have the suppression effect of storms on equatorial plasma bubble irregularities. Asymmetry between two equinoxes was observed. The occurrence rate of equatorial plasma bubble irregularities in the March equinox was much larger than that of the September equinox. The occurrence probability of equatorial plasma bubbles for March Equinox was maximum with 45,1% at 17 UT while the occurrence probability of equatorial plasma bubbles for September Equinox was maximum with 11,5% at 16 UT. The enhancements and reductions in the latitudinal gradient of VTEC show similar behaviors with the occurrence of equatorial plasma irregularities. The EIA during postsunset hours contributes significantly to the occurrence of equatorial plasma bubbles irregularities.
Detection of Stylometric Writeprint From the Turkish Texts
(Ieee, 2020) Canbay, Pelin; Sever, Hayri; Sezer, Ebru Akcapinar; Sever, Hayri; Bilgisayar Mühendisliği
Authorship attribution studies aim to extract information about the author by analyzing the data in the text form. With the increase of anonymous authors in digital environments, the need for these works is increasing day by day. Although there exists lots of studies focuse on stylometric writeprint detection in different languages using different attributes, there is no standard feature set and detection algorithm to be evaluated in these studies. Giving priority to Turkish texts, in this study, which features are more distinctive for determining stylistic writeprint of text, and which methods will contribute to increase the success to be achieved are shown with experimental studies.
Citation - WoS: 228
Citation - Scopus: 297
A Comprehensive Survey on Recent Metaheuristics for Feature Selection
(Elsevier, 2022) Dokeroglu, Tansel; Deniz, Ayca; Kiziloz, Hakan Ezgi
Feature selection has become an indispensable machine learning process for data preprocessing due to the ever-increasing sizes in actual data. There have been many solution methods proposed for feature selection since the 1970s. For the last two decades, we have witnessed the superiority of metaheuristic feature selection algorithms, and tens of new ones are being proposed every year. This survey focuses on the most outstanding recent metaheuristic feature selection algorithms of the last two decades in terms of their performance in exploration/exploitation operators, selection methods, transfer functions, fitness value evaluations, and parameter setting techniques. Current challenges of the metaheuristic feature selection algorithms and possible future research topics are examined and brought to the attention of the researchers as well.
Sınıflandırmada Küçük ve Dengesiz Veri Kümesi Problemi
(2019) Par, Öznur Esra; Akçapınar Sezer, Ebru; Sever, Hayri
Verilerinin sınıflandırılması, veri kümesinin küçük ve dengesiz olması durumunda zorlaşmakta ve sınıflama performansını direkt etkilemektedir. Veri setinin küçük olması ve/veya sınıflar arasında dengesizlik olması veri madenciliğinde büyük bir sorun haline gelmiştir. Sınıflama algoritmaları, veri setlerinin yeterli büyüklüğe sahip, dengeli olduğu varsayımı üzerine geliştirilmiştir. Bu algoritmaların çoğu, azınlık sınıfındaki örnekleri göz ardı ederken veya yanlış sınıflandırırken, çoğunluk sınıfa odaklanır. Medikal veri madenciliğinde bazı kısıtlardan dolayı küçük ve dengesiz veri seti problemi ile sıklıkla karşılaşılmaktadır. Çalışma kapsamında erişime açık hepatit veri seti, küçük veri setlerine bölünmüş, oluşturulan her bir veri seti uzaklık tabanlı yöntemlerle çoğaltılmıştır. Çoğaltılan veri setleri dört farklı makine öğrenmesi algoritması (Yapay Sinir Ağları, Destek Vektör Makineleri, Naive Bayes ve Karar Ağacı) kullanılarak sınıflandırılmış, elde edilen sınıflama sonuçları karşılaştırılmıştır.
Big Data Reduction and Visualization Using the K-Means Algorithm
(2022) Akyol, Hakan; Kızılduman, Hale Sema; Dökeroğlu, Tansel
A huge amount of data is being produced every day in our era. In addition to high-performance processing approaches, efficiently visualizing this quantity of data (up to Terabytes) remains a major difficulty. In this study, we use the well-known clustering method K-means as a data reduction strategy that keeps the visual quality of the provided huge data as high as possible. The centroids of the dataset are used to display the distribution properties of data in a straightforward manner. Our data comes from a recent Kaggle big data set (Click Through Rate), and it is displayed using Box plots on reduced datasets, compared to the original plots. It is discovered that K-means is an effective strategy for reducing the amount of huge data in order to view the original data without sacrificing its distribution information quality
Citation - WoS: 2
Citation - Scopus: 4
An Intelligent System for Detecting Mediterranean Fruit Fly [Medfly; Ceratitis Capitata (Wiedemann)]
(Pagepress Publ, 2022) Eyyuboglu, Halil Tanyer; Sari, Filiz; Uzun, Yusuf; Tolun, Mehmet Resit
Nowadays, the most critical agriculture-related problem is the harm caused to fruit, vegetable, nut, and flower crops by harmful pests, particularly the Mediterranean fruit fly, Ceratitis capitata, named Medfly. Medfly's existence in agricultural fields must be monitored systematically for effective combat against it. Special traps are utilised in the field to catch Medflies which will reveal their presence and applying pesticides at the right time will help reduce their population. A technologically supported automated remote monitoring system should eliminate frequent site visits as a more economical solution. This paper develops a deep learning system that can detect Medfly images on a picture and count their numbers. A particular trap equipped with an integrated camera that can take photos of the sticky band where Medflies are caught daily is utilised. Obtained pictures are then transmitted by an electronic circuit containing a SIM card to the central server where the object detection algorithm runs. This study employs a faster region-based convolutional neural network (Faster R-CNN) model in identifying trapped Medflies. When Medflies or other insects stick on the trap's sticky band, they spend extraordinary effort trying to release themselves in a panic until they die. Therefore, their shape is badly distorted as their bodies, wings, and legs are buckled. The challenge is that the deep learning system should detect these Medflies of distorted shape with high accuracy. Therefore, it is crucial to utilise pictures containing trapped Medfly images with distorted shapes for training and validation. In this paper, the success rate in identifying Medflies when other insects are also present is approximately 94%, achieved by the deep learning system training process, owing to the considerable amount of purpose-specific photographic data. This rate may be seen as quite favourable when compared to the success rates provided in the literature.
Yeşil BHT Bilgi ve Haberleşme Teknolojileri Akademisyen ve Uygulayıcılar Açısından Bir İnceleme
(2011) Akba, Fırat; Medeni, İhsan Tolga; Medeni, Tunç Durmuş; Tolun, Mehmet Reşit; Öztürk, Mehmet
Citation - Scopus: 1
Component-Based Project Estimation Issues for Recursive Development
(Springer, 2008) Altunel, Yusuf; Tolun, Mehmet R.
In this paper we investigated the component-based specific issues that might affect project cost estimation. Component-based software development changes the style of software production. With component-based approach the software is developed as the composition of reusable software components. Each component production process must be treated as a stand-alone software project, which needs individual task of management. A typical pure component-based development can be considered as decomposition/integration activities successively applied at different levels and therefore results in recursive style of development. We analyzed and presented our results of studies on the component-based software development estimation issues from recursive point of view.
Müfredat Tabanlı Ders Çizelgeleme Problemi için Yeni Bir Açgözlü Algoritma
(2023) Batuhan,; Say, Bilge; Dokeroglu, Tansel
Bu çalışma, iyi bilinen Müfredat Tabanlı Ders Çizelgeleme Problemini optimize etmek için yeni bir açgözlü algoritmayı açıklamaktadır. Açgözlü algoritmalar, en iyi çözümü bulmak için yürütülmesi uzun zaman alan kaba kuvvet ve evrimsel algoritmalara iyi bir alternatiftir. Birçok açgözlü algoritmanın yaptığı gibi tek bir buluşsal yöntem kullanmak yerine, aynı problem örneğine 120 yeni buluşsal yöntem tanımlıyor ve uyguluyoruz. Dersleri müsait odalara atamak için, önerilen açgözlü algoritmamız En Büyük-İlk, En Küçük-İlk, En Uygun, Önce Ortalama Ağırlık ve En Yüksek Kullanılamaz ders-ilk buluşsal yöntemlerini kullanır. İkinci Uluslararası Zaman Çizelgesi Yarışması'nın (ITC-2007) kıyaslama setinden 21 problem örneği üzerinde kapsamlı deneyler gerçekleştirilir. Önemli ölçüde azaltılmış yumuşak kısıtlama değerlerine sahip 18 problem için, önerilen açgözlü algoritma sıfır sabit kısıtlama ihlali (uygulanabilir çözümler) rapor edebilir. Önerilen algoritma, performans açısından son teknoloji ürünü açgözlü buluşsal yöntemleri geride bırakıyor.
Citation - WoS: 3
Citation - Scopus: 5
A New Robust Harris Hawk Optimization Algorithm for Large Quadratic Assignment Problems
(Springer London Ltd, 2023) Dokeroglu, Tansel; Ozdemir, Yavuz Selim
Harris Hawk optimization (HHO) is a new robust metaheuristic algorithm proposed for the solution of large intractable combinatorial optimization problems. The hawks are cooperative birds and use many intelligent hunting techniques. This study proposes new HHO algorithms for solving the well-known quadratic assignment problem (QAP). Large instances of the QAP have not been solved exactly yet. We implement HHO algorithms with robust tabu search (HHO-RTS) and introduce new operators that simulate the actions of hawks. We also developed an island parallel version of the HHO-RTS algorithm using the message passing interface. We verify the performance of our proposed algorithms on the QAPLIB benchmark library. One hundred and twenty-five of 135 problems are solved optimally, and the average deviation of all the problems is observed to be 0.020%. The HHO-RTS algorithm is a robust algorithm compared to recent studies in the literature.
Citation - WoS: 2
Citation - Scopus: 4
Exploring Mooc Learners' Behavioural Patterns Considering Age, Gender and Number of Course Enrolments: Insights for Improving Educational Opportunities
(int Council Open & Distance Education, 2024) Cagiltay, Nergiz ercil; Toker, Sacip; Cagiltay, Kursat
Massive Open Online Courses (MOOCs) now offer a variety of options for everyone to obtain a high -quality education. The purpose of this study is to better understand the behaviours of MOOC learners and provide some insights for taking actions that benefit larger learner groups. Accordingly, 2,288,559 learners' behaviours on 174 MITx courses were analysed. The results show that MOOCs are more attractive to the elderly, male, and highly educated groups of learners. Learners' performance improves as they register for more courses and improve their skills and experiences on MOOCs. The findings suggest that, in the long run, learners' adaptation to MOOCs will significantly improve the potential benefits of the MOOCs. Hence, MOOCs should continue by better understanding their learners and providing alternative instructional designs by considering different learner groups. MOOC providers' decision -makers may take these findings into account when making operational decisions.
Hierarchical SNR Scalable Video Coding with Adaptive Quantization for Reduced Drift Error
(2015) Choupani, Roya; Wong, Stephan; Tolun, Mehmet
In video coding, dependencies between frames are being exploited to achieve compression by only coding the differences. This dependency can potentially lead to decoding inaccuracies when there is a communication error, or a deliberate quality reduction due to reduced network or receiver capabilities. The dependency can start at the reference frame and progress through a chain of dependent frames within a group of pictures (GOP) resulting in the so-called drift error. Scalable video coding schemes should deal with such drift errors while maximizing the delivered video quality. In this paper, we present a multi-layer hierarchical structure for scalable video coding capable of reducing the drift error. Moreover, we propose an optimization to adaptively determine the quantization step size for the base and enhancement layers. In addition, we address the trade-off between the drift error and the coding efficiency. The improvements in terms of average PSNR values when one frame in a GOP is lost are 3.70(dB) when only the base layer is delivered, and 4.78(dB) when both the base and the enhancement layers are delivered. The improvements in presence of burst errors are 3.52(dB) when only the base layer is delivered, and 4.50(dB) when both base and enhancement layers are delivered.
Citation - WoS: 1
Citation - Scopus: 4
A Comparative Evaluation of Popular Search Engines on Finding Turkish Documents for A Specific Time Period
(Univ Osijek, Tech Fac, 2017) Gorur, Abdul Kadir; Bitirim, Yiltan
This study evaluates the popular search engines, Google, Yahoo, Bing, and Ask, on finding Turkish documents by comparing their current performances with their performances measured six years ago. Furthermore, the study reveals the current information retrieval effectiveness of the search engines. First of all, the Turkish queries were run on the search engines separately. Each retrieved document was classified and precision ratios were calculated at various cut-off points for each query and engine pair. Afterwards, these ratios were compared with the six years ago ratios for the evaluations. Besides the descriptive statistics, Mann-Whitney U and Kruskal-Wallis H statistical tests were used in order to find out statistically significant differences. All search engines, except Google, have better performance today. Bing has the most increased performance compared to six years ago. Nowadays: Yahoo has the highest mean precision ratios at various cut-off points; all search engines have their highest mean precision ratios at cut-off point 5; dead links were encountered in Google, Bing, and Ask; and repeated documents were encountered in Google and Yahoo.
Türk Beyin Cerrahlarının Teknolojiye Ulaşım İmkanları
(2018) Çağıltay, Nergiz
Otomatik Konuşma Tanıma Sistemlerinde Kullanılan Gerçek Metin Verisinde Biçimbilimsel-Sözdizimsel Hataların Tespiti ve Düzeltmesi
(2019) Polat, Hüseyin; Sever, Hayri; Oyucu, Saadin; Tekbaş, Şükran
Türkçe Otomatik Konuşma Tanıma (ASR: Automatic Speech Recognition) sistemlerinde kullanılan akustik model gürbüz bir dil modeli ile desteklenmediği durumlarda kelime hata oranı yüksek çıkmaktadır. İyi dizayn edilmiş bir dil modeli ile akustik modelin birlikte ASR’de kullanılması kelime hata oranını düşürmektedir. ASR için gerekli dil modelinin eğitiminde düz metin verisi kullanılmaktadır. Kullanılan metin verisinin doğruluğu ASR modellerinin eğitimi için oldukça önemlidir. Bu çalışmada, doğal dil işlemeye dayalı bir yöntem kullanılarak Türkçe ASR sisteminin eğitilmesinde kullanılan metin verisi içerisindeki yazım hatalarının tespiti ve düzeltilmesi gerçekleştirilmiştir. Öncelikle metin verisi içerisinde dil bilgisel olarak yanlış yazılmış olan kelimeler bulunmuştur. Bir kelimedeki karakter eksikliği, karakter fazlalığı, karakterlerin yer değiştirmesi veya karakteri yanlış yazılmış olan kelimeler hatalı olarak kabul edilmiştir. Metin verisi içerisinde hatalı olarak kabul edilen kelimeler morfolojik analiz ile tespit edilmiştir. Yanlış kelimelerin yerine atanacak olan kelimeler belirlenmiştir. Yanlış yazılmış olan kelimeler doğru kelimeler ile değiştirilmiştir. Gerçekleştirilen çalışma hatalı kelimeleri tespit etme ve doğru kelimeler ile yer değiştirme işleminde %93 oranında başarı göstermiştir.
Sessizliğin Kaldırılması ve Konuşmanın Parçalara Ayrılması İşleminin Türkçe Otomatik Konuşma Tanıma Üzerindeki Etkisi
(2020) Sever, Hayri; Polat, Huseyin; Oyucu, Saadin
Otomatik Konuşma Tanıma sistemleri temel olarak akustik bilgiden faydalanılarak geliştirilmektedir. Akustikbilgiden fonem bilgisinin elde edilmesi için eşleştirilmiş konuşma ve metin verileri kullanılmaktadır. Bu verilerile eğitilen akustik modeller gerçek hayattaki bütün akustik bilgiyi modelleyememektedir. Bu nedenle belirli önişlemlerin yapılması ve otomatik konuşma tanıma sistemlerinin başarımını düşürecek akustik bilgilerin ortadankaldırılması gerekmektedir. Bu çalışmada konuşma içerisinde geçen sessizliklerin kaldırılması için bir yöntemönerilmiştir. Önerilen yöntemin amacı sessizlik bilgisinin ortadan kaldırılması ve akustik bilgide uzunbağımlılıklar sağlayan konuşmaların parçalara ayrılmasıdır. Geliştirilen yöntemin sonunda elde edilen sessizlikiçermeyen ve parçalara ayrılan konuşma bilgisi bir Türkçe Otomatik Konuşma Tanıma sistemine girdi olarakverilmiştir. Otomatik Konuşma Tanıma sisteminin çıkışında sisteme giriş olarak verilen konuşma parçalarınakarşılık gelen metinler birleştirilerek sunulmuştur. Gerçekleştirilen deneylerde sessizliğin kaldırılması vekonuşmanın parçalara ayrılması işleminin Otomatik Konuşma Tanıma sistemlerinin başarımını artırdığıgörülmüştür.
Citation - WoS: 5
Citation - Scopus: 8
Predicting the Severity of Covid-19 Patients Using a Multi-Threaded Evolutionary Feature Selection Algorithm
(Wiley, 2022) Kiziloz, Hakan Ezgi; Sevinc, Ender; Dokeroglu, Tansel; Deniz, Ayca
The COVID-19 pandemic has huge effects on the global community and an extreme burden on health systems. There are more than 185 million confirmed cases and 4 million deaths as of July 2021. Besides, the exponential rise in COVID-19 cases requires a quick prediction of the patients' severity for better treatment. In this study, we propose a Multi-threaded Genetic feature selection algorithm combined with Extreme Learning Machines (MG-ELM) to predict the severity level of the COVID-19 patients. We conduct a set of experiments on a recently published real-world dataset. We reprocess the dataset via feature construction to improve the learning performance of the algorithm. Upon comprehensive experiments, we report the most impactful features and symptoms for predicting the patients' severity level. Moreover, we investigate the effects of multi-threaded implementation with statistical analysis. In order to verify the efficiency of MG-ELM, we compare our results with traditional and state-of-the-art techniques. The proposed algorithm outperforms other algorithms in terms of prediction accuracy.

Browse

Browsing Yazılım Mühendisliği Bölümü by Issue Date