Sürdürülebilir Kalkınma Amaçları
Yazarlar
- Ayçin, Ejder(2)
- Kaya, Sema Kayapinar(2)
- Bulut, Zeki Atil(1)
- Demir, Yakup(1)
- Dogan, Onur(1)
- Entani, T.(1)
- Gunay, Mihriban(1)
- Huynh, V.(1)
- Hüseyin Bolat(1)
- Inuiguchi, M.(1)
- Jeenanunta, C.(1)
- Karlıoğlu Kılıç, Nurgül(1)
- Mikhov, Mikho(1)
- Navarro-Arribas, Guillermo(1)
- Paksoy, Mehmet Yavuz(1)
- Pamucar, Dragan(1)
- Torra, Vicenç(1)
- Tuncalı Yaman, Tutku(1)
- Yenradee, P.(1)
- Yildirim, Ozal(1)
- Yordanov, Nikolay(1)
- Yılmaz Dağdeviren, Rüya(1)
- Zhilevski, Marin(1)
Yayınlar
- MakaleAçık Erişim2026Scopus SJRQ2Hybrid Supervised Classification and Deep Embedding–Based Profiling Framework for Electricity Consumption Analysis
Gunay, Mihriban; Yildirim, Ozal; Demir, Yakup; Zhilevski, Marin; Mikhov, Mikho; Yordanov, Nikolay
This study proposes a hybrid deep learning framework that integrates supervised classification and unsupervised profiling for electricity consumption analysis. In the supervised phase, a one-dimensional Convolutional Neural Network combined with Long Short-Term Memory (1D CNN–LSTM) architecture is developed to classify daily load patterns. The performance of the proposed model is compared with traditional machine learning and deep learning approaches, including Support Vector Machine (SVM), k-Nearest Neighbors (KNN), a standalone Long Short-Term Memory (LSTM) model, a Transformer-based model, and a standalone 1D CNN model. Experimental results on the Precon house dataset and the CU-BEMS dataset demonstrate that the proposed hybrid architecture outperforms the benchmark models, achieving classification accuracies of 87.59% and 86.40%, respectively. In the unsupervised phase, the trained CNN–LSTM encoder is utilized as a deep feature extractor. The resulting 32-dimensional latent embeddings are clustered using K-Means, Gaussian Mixture Model (GMM), Agglomerative, Spectral, and Ensemble methods. Clustering robustness is evaluated through bootstrap-based stability analysis using the Adjusted Rand Index (ARI) and the Normalized Mutual Information (NMI). The results demonstrate stable and interpretable electricity consumption profiles, particularly in the residential dataset, where near-perfect clustering stability is observed for K-Means. The proposed framework provides both improved classification performance and robust consumption profiling based on deep embedding, offering a practical tool for energy management.
AtıfScopus1OpenAlex1
İncele - MakaleYalnız künye2023Scopus SJRQ3Evaluation of social factors within the circular economy concept for European countries
Kaya, Sema Kayapinar; Ayçin, Ejder; Pamucar, Dragan
The circular economy (CE) is a rapidly growing theme, particularly in the European Union (EU), that encourages the responsible and circular use of resources in the field of contributing to long-term development. The environmental and economic magnitude of development are frequently debated in the subject of CE whereas, social aspects have been only rarely and sporadically combined into the CE. Therefore, this study involves a multidisciplinary holistic framework of clustering methods and Multi-Criteria Decision Making (MCDM) for evaluating the CE paradigm in relation to EU countries' social growth. In terms of the social impact of CE strategies, the k-means cluster analysis was used to group the 27 EU members with similar social impact levels. Subsequently, a novel integrated CRITIC (The criteria importance through intercriteria correlation) and MEREC (Method based on the Removal Effects of Criteria) methods were employed for determining the weights of social indicators in order to balance the two weighting methods. In order to select the best country in each cluster, a novel extension of the hybrid MARCOS (Measurement of Alternatives and Ranking according to Compromise Solution) method combined with Power averaging and Heronian operator was developed. According to final solutions, Western European countries in the first cluster have the lowest unemployment and corruption rates, with the Netherlands having the best performance in this cluster. The second cluster includes countries with the lowest employment rates following university graduates between the ages of 20 and 64. Accordingly, Croatia is the best social performance in this cluster. Countries with the highest income distribution and unemployment rate are in the 3rd cluster, the best country in this group is Lithuania. Finally, the results obtained from the innovative MCDM methods have validated in order to demonstrate the proposed methodology's applicability.
AtıfScopus30WoS23OpenAlex42
İncele - Yüksek Lisans TeziAçık Erişim2022Makine öğrenme yöntemleri ile Kahramanmaraş ilindeki trafik kazası sonuçlarının analizi ve tahmin edilmesi
Hüseyin Bolat
Trafik kazaları insan yaralanmaları ve ölümlerinin en büyük nedenlerindendir. Dünya üzerinde her yıl 1 milyonun üzerinde ölümlü, 50 milyona yakın yaralanmalı trafik kazası meydana gelmektedir. Bu kazalarda ortalama her 2 saniyede bir trafik kazası yaşanırken her 50 saniyede bir ölümcül kaza ile sonuçlanmaktadır. Bu çalışmada, 2017 ile 2020 yılları arasında Kahramanmaraş il merkezi ve ilçelerinde gerçekleşen 28 adet öznitelikten oluşan 7929 adet trafik kazası WEKA aracında makine öğrenme teknikleri ile analiz edildi. Trafik kazalarının analizinde sınıflandırma, kümeleme, birliktelik kuralı ve görsel karşılaştırma yöntemleri kullanılmaktadır. Sınıflandırma yönteminde Naive Bayes, Destek Vektör Makinesi, K En Yakın Komşu, PART, AdaboostM1, J48 ve Rastgele Orman algoritmaları kullanıldı. Kümeleme yönteminde SimpleKMeans ve EM kümeleme algoritmaları kullanıldı. Birliktelik yönteminde Apriori algoritmaları kullanıldı. Sonuç olarak trafik kazalarının tahmininde sınıflandırma algoritmaların performans ölçütlerine bakıldığında karar ağaçlarından J48 algoritması en başarılı sonuçları elde ederken, NB algoritmasının başarı düzeyi kötü olmasa da diğer algoritmalara oranla daha düşük kalmıştır. Tüm yöntemlerin ortak sonucu gelecekte yaşanabilecek ölümle ve yaralanmalı potansiyel kazaların daha çok sabah saat 04:00 ile 08:00 ve akşam saat 16:00 ile 20:00 arasında, Kahramanmaraş il merkezi ve Onikişubat ilçesinde yerleşim yeri içerisinde, Göksun ilçesindeki yerleşim yeri dışındaki eğimli ve virajlı yollarda ve Pazarcık ilçesindeki yerleşim yeri dışındaki bölünmüş yollarda yaşanma riskinin yüksek olduğu belirlenmiştir. Ayrıca meydana gelen kazalarda sürücülerin trafik ve hız kurallarına uymadığı, trafik lambası, trafik levhası, yaya yolu, emniyet şeridi ve banket gibi yol, çevre ve güvenlik ekipmanlarının yetersiz kaldığı belirlenmiştir.
- MakaleAçık Erişim2021Scopus SJRQ2Pollen morphology of eight endemic Inula L. (Asteraceae) species in Turkey
Karlıoğlu Kılıç, Nurgül; Yılmaz Dağdeviren, Rüya; Paksoy, Mehmet Yavuz; Tuncalı Yaman, Tutku
The pollen morphology of eight endemic species of the genus Inula L. (Asteraceae)–I. helenium subsp. orgyalis (Boiss.) Grierson, I. helenium subsp. vanensis Grierson, I. macrocephala Boiss. & Kotschy ex Boiss., I. discoidea Boiss., I. sarana Boiss., I. fragilis Boiss. & Hausskn., I. anatolica Boiss. and I. sechmenii Hartvig & Strid–distributed in Turkey was studied with light microscopy and scanning electron microscopy. We found that the pollen of these Inula taxa are prolate-spheroidal in shape (except in Inula discoidea), with the polar axis ranging from 19.18 to 33.14 µm and the equatorial diameter ranging from 18.54 to 32.92 µm. The pollen grains are tricolporate and their ornamentation is echinate. The exine thickness ranges from 0.87 to 2.07 µm. As this study demonstrates, results of Principal Component Analysis are revealed that some of the pollen characters, such as spine length and width, are significant morphological characters for taxonomic discrimination of the eight endemic Inula taxa. According to cluster analysis, two main groups of examined taxa are constituted.
AtıfScopus9WoS9OpenAlex10İncele - BildiriYalnız künye2020Scopus SJRQ3Dynamic Features Spaces and Machine Learning: Open Problems and Synthetic Data Sets
Kaya, Sema Kayapinar; Navarro-Arribas, Guillermo; Torra, Vicenç; Huynh, V.; Entani, T.; Jeenanunta, C.; Inuiguchi, M.; Yenradee, P.
Dynamic feature spaces appear when different records or instances in databases are defined in terms of different features. This is in contrast with usual (static) feature spaces in standard databases, where the schema of the database is known and fixed. Then, all records in the database have the same set of variables, attributes or features. Dynamic feature mining algorithms are to extract knowledge from data on dynamic feature spaces. As an example, spam detection methods have been developed from a dynamic feature space perspective. Words are taken as features and new words appearing in new emails are, therefore, considered new features. In this case, the problem of spam detection is represented as a classification problem (a supervised machine learning problem). The relevance of dynamic feature spaces is increasing. The large amounts of data currently available or received by systems are not necessarily described using the same feature spaces. This is the case of distributed databases with data about customers, providers, etc. Industry 4.0, Internet of Things, and RFIDs are and will be a source of data in dynamic feature spaces. New sensors added in an industrial environment, new devices connected into a smart home, new types of analysis and new types of sensors in healthcare, all are examples of dynamic feature spaces. Machine learning algorithms are needed to deal with these type of scenarios. In this paper we motivate the interest for dynamic feature mining, we give some examples of scenarios where these techniques are needed, we review some of the existing solutions and its relationship with other areas of machine learning and data mining (e.g., incremental learning, concept drift, topic modeling), we discuss some open problems, and we discuss synthetic data generation for this type of problem.
AtıfScopus4WoS3OpenAlex3İncele - MakaleYalnız künye2018CUSTOMER SEGMENTATION BY USING RFM MODEL AND CLUSTERING METHODS: A CASE STUDY IN RETAIL INDUSTRY
Dogan, Onur; Ayçin, Ejder; Bulut, Zeki Atil
Companies need to understand the customers' data better in all aspects. Detecting similarities and differences among customers, predicting their behaviors, proposing better options and opportunities to customers became very important for customer-company engagement. Segmenting the customers according to their data became vital in this context. RFM (recency, frequency and monetary) values have been used for many years to identify which customers valuable for the company, which customers need promotional activities, etc. Data-mining tools and techniques widely have been used by organizations and individuals to analysis their stored data. Clustering, which one of the tasks of data mining has been used to group people, objects, etc. In this paper we propose two different clustering models to segment 700032 customers by considering their RFM values. We detected that the current customer segmentation which built by just considering customers' expense is not sufficient. Hence, models that recommended in this research are expected to provide better customer understanding, well-designed strategies, and more efficient decisions.
AtıfWoS30İncele