Metin Turan

Metin Turan

Istanbul Ticaret University, Istanbul, İstanbul, Turkey (Türkiye)
orcid

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 January 2021
Verified
Conference Paper
Closed access
SÖSüley Man Özdemirci
Süley Man Özdemirci
MTMetin Turan
Metin Turan
Topic modeling has numerous applications like text categorization, topic clustering, document tagging, feature extraction on wide document collections. In this study, practical exploration method of topic modeling of Latent Dirichlet Allocation, transformers based machine learning method Bidirectional Encoder Representations from Transformers and Term Frequency - Inverse Document Frequency method were applied to the document set separately. It includes sport and education articles collected from internet by graduate students, 801 number totally. The purpose of this study is to observe which method best suits to the topic modeling and if possible in order to increase the accuracy rate via ensemble of these methods. As a result of this study, it was observed that even it has some disadvantages, BERT classified the documents with the correct topic with an average of %92.6 success ratio, overwhelming the others.

1 November 2007
Verified
Conference Paper
Closed access
MTMetin Turan
Metin Turan
ZCZehra Cataltepe
Zehra Cataltepe
During the last two decades research on software engineering is concentrated on quality. The best approach to quality evaluation goes through determining well-defined metrics on software properties. One such property is module complexity, which is a view of the software that is related to how easily it can be modified. There has been work on constructing a metrics domain which measures the module complexity. Generally, PCA (Principal Component Analysis) is used for defining principal metrics in the domain. Since there are usually no labels for the software data, an unsupervised dimensionality reduction technique, such as PCA needs to be used for determining the most important metrics. In this study, we use the clustering similarity obtained when a certain subset of metrics and when the whole set of metrics are used, to determine the most important metrics. We measure the relative difference/similarity between clusterings using three different indices, namely Rand, Jaccard and Fowlkes-Mallow. We use both backward feature selection and PCA for dimensionality reduction. On the publicly available NASA data, we find out that instead of the whole set of 42 metrics, using only 15 dimensions, we get almost the same clustering performance. Therefore, instead of the whole set of software metrics, a smaller number of them could be used to evaluate the software quality.

28 November 2024 in Data and Information Management
Verified
Journal Article
Open access
ATAhmet Toprak
Ahmet Toprak
MTMetin Turan
Metin Turan

28 June 2020 in Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji
Verified
Journal Article
Closed access
PDF file
RKRabia Kontuk
Rabia Kontuk
MTMetin Turan
Metin Turan
Teknoloji ile birlikte hayatımıza birçok yeni kavram girmiştir. Bu yeniliklerden biri de Doğal Dil İşleme (DDİ) kavramıdır. Bu kavram beraberinde birçok çalışma alanı getirdiği için Doğal Dil İşleme ile yapılan çalışmalar gün geçtikçe artmaktadır. Bu çalışmada, Doğal Dil işleme kullanılarak elektronik ortamlardaki haberlerin yaş gruplarına göre etiketlenmesi amaçlanmıştır. Haber sitelerinden toplanan haber veri setinin eğitim amaçlı seçilmiş olanları, NLP Zemberek Kütüphanesi kullanılarak Python dili ile işlenmiş, Havighurst’ün “Gelişim Kuramı” nın güncel duruma adapte edilmiş Çocukluk, Ergenlik ve Yetişkinlik yaş gruplarını temsil edebilecek kelime sözlüğü oluşturulmuştur (her kelimenin hangi yaş grubuna uygun olduğu). Daha sonra, bu sözlük kullanılarak haber veri setinin test amaçlı seçilmiş olanlarının sınıflarını belirlemek üzere bir sınıflandırıcı önerilmiştir. Testler sonucunda, geliştirilen sözlüğün 0.70 oranında doğru sınıfı tespit edebildiği görülmüştür.

11 June 2021
Verified
Conference Paper
Closed access
MIMert İleri
Mert İleri
MTMetin Turan
Metin Turan
In the last decade, enormous data are being shared throughout the world. In many of today’s big data world, the companies are trying to use some sentiment or emotion analysis techniques to analyze their customer moods and improve their efficiencies according to sentiments. As a different application we focused on the sentiment analysis of closed places in this research. It requires low noise environments obviously. Otherwise, system may be affected by distortion, and it may be contradiction for multiple sentiments. In this regard, an artificial neural network using meaningful voice features are proposed. Ryerson Audio Visual Database of Emotional Speech and Song (RAVDESS) dataset was used in this research. Normalization was applied to data. The artificial neural network was fed by training data and a classifier model was created. Estimation was made using the test data part and it was seen that accuracy of model is about 85%.

1 November 2019
Verified
Conference Paper
Closed access
ATAhmet Toprak
Ahmet Toprak
MTMetin Turan
Metin Turan