Türkçede Yüksek Doğruluklu Duygu Tespiti için Makine Öğrenmesi Sınıflandırıcılarının Optimizasyonu

Öz

Duygu analizi, metinsel verilerden görüşleri ortaya çıkarmak için yaygın olarak kullanılmaktadır; ancak Türkçe gibi morfolojik açıdan zengin dillerdeki uygulamaları zorluğunu korumaktadır. Bu çalışma, birleşik bir deneysel çerçeve altında, ikili (binary) Türkçe duygu analizi için klasik makine öğrenmesi sınıflandırıcılarının ve topluluk öğrenme (ensemble learning) stratejilerinin optimizasyonunu incelemektedir. Dengeli bir kullanıcı yorumları veri seti üzerinde, doğrusal destek vektör makinesi sınıflandırıcısı (LSVMC), çok terimli Naïve Bayes (MNB) ve lojistik regresyon (LR) dahil olmak üzere çeşitli makine öğrenmesi modelleri eğitilmiştir. Bu modellerin çıktıları, çoğunluk oylaması (MVEL) ve yığınlama (SEL) gibi topluluk öğrenme yöntemleri kullanılarak birleştirilmiştir. Sonuçlar, SEL sınıflandırıcısının artan hesaplama karmaşıklığı pahasına %92,80 doğruluk oranına ulaşarak incelenen tüm modellerden daha iyi performans gösterdiğini ortaya koymaktadır. İncelenen bireysel modeller arasında en yüksek doğruluk oranlarını LSVMC (%92), MNB (%92) ve LR (%92) sergilemiştir. Çalışma, derin öğrenme veya transformatör tabanlı mimariler aracılığıyla literatürdeki en ileri düzey (state-of-theart) performansı geride bırakmayı hedeflemese de, sonuçlar, optimize edilmiş klasik modellerin Türkçe duygu analizinde halen oldukça başarılı olduğunu ve performanslarının topluluk öğrenme mekanizmalarıyla daha da geliştirilebileceğini kanıtlamaktadır

Anahtar Kelimeler

• Makine Öğrenmesi
• Topluluk Öğrenmesi
• Türkçe Dil İşleme
• Türkçe Duygu Analizi

Destekleyen Kurum

Hasan Kalyoncu Üniversitesi / Türkiye

Etik Beyan

Bu araştırma, yazarlar tarafından insan katılımcılar veya hayvanlar üzerinde gerçekleştirilen herhangi bir çalışmayı içermemektedir. Bu çalışmada analiz edilen veri seti, Hacettepe Üniversitesi/Türkiye resmi web sitesinden akademik araştırma izni alınarak elde edilmiş olup, bu nedenle ek bir etik onayına ihtiyaç duyulmamıştır.

Teşekkür

Bilimsel araştırma derslerinde bana öğrettiği araştırma becerileri için Dr. İbrahim Halil Değer'e teşekkür etmek istiyorum. Ayrıca, bu araştırmayı tamamlamam için gerekli bilgi ve uzmanlığı bana sağlayan herkese teşekkür etmek istiyorum.

OPTIMIZING MACHINE LEARNING CLASSIFIERS FOR HIGH-ACCURACY SENTIMENT DETECTION IN THE TURKISH LANGUAGE

Öz

Sentiment analysis is widely used to extract opinions from textual data; however, its application to morphologically rich languages such as Turkish remains challenging. This study investigates the optimization of classical machine learning classifiers and ensemble learning strategies for binary Turkish sentiment analysis under a unified experimental framework. Several ML models are trained on a balanced dataset of user reviews, including Linear Support Vector Machine Classifier (LSVMC), Multinomial Naïve Bayes (MNB), and Logistic Regression (LR). Their outputs were further combined using Ensemble Learning (EL) models, namely Majority Voting (MVEL) and Stacking (SEL). Results demonstrate that the SEL Classifier outperforms all examined models, achieving 92.80% accuracy at the cost of increased computational complexity. Among the examined individual models, LSVMC (92%), MNB (92%), and LR (92%) had the best accuracy. While the study does not aim to achieve state-of-the-art performance with deep or transformer-based architectures, the results demonstrate that optimized classical models remain highly effective in Turkish SA, and their accuracy can be further improved with EL mechanisms.

Anahtar Kelimeler

• Machine Learning
• Ensemble Learning
• Turkish Language Processing
• Turkish Sentiment Analysis

Destekleyen Kurum

Hasan Kalyoncu University / Turkey

Etik Beyan

This research did not involve any studies with human participants or animals performed by the authors. The dataset analyzed in this study was obtained with permission for academic research from Hacettepe University/Türkiye official website, and therefore no further ethical approval was required.

Teşekkür

I would like to thank Dr. İbrahim Halil Değer for the research skills he taught me in the scientific research classes. I would also like to thank everyone who provided me with the necessary knowledge and expertise to complete this research.

Kaynakça

1. Armaeni, P. P., Wiguna, I. K. A. G., & Parwita, W. G. S. (2024). Sentiment analysis of YouTube comments on the closure of TikTok Shop using Naïve Bayes and Decision Tree method comparison. Jurnal Galaksi, 1(2), 70-80.
2. Aydoğan, M., & Kocaman, V. (2023). TRSAv1: A new benchmark dataset for classifying user reviews on Turkish e-commerce websites. Journal of Information Science, 49(6), 1711-1725.
3. Başarslan, M. S., & Kayaalp, F. (2023). Sentiment analysis with ensemble and machine learning methods in multi-domain datasets. Turkish Journal of Engineering, 7(2), 141-148.
4. Danyal, M. M., Khan, S. S., Khan, M., Ullah, S., Ghaffar, M. B., & Khan, W. (2024). Sentiment analysis of movie reviews based on NB approaches using TF–IDF and count vectorizer. Social network analysis and mining, 14(1), 87.
5. Demircan, M., Seller, A., Abut, F., & Akay, M. F. (2021). Developing Turkish sentiment analysis models using machine learning and e-commerce data. International Journal of Cognitive Computing in Engineering, 2, 202-207.
6. Demirci, G. M., Keskin, Ş. R., & Doğan, G. (2019, December). Sentiment analysis in Turkish with deep learning. In 2019 IEEE international conference on big data (big data) (pp. 2215-2221). IEEE.
7. Erşahin, B., Aktaş, Ö., Kilinç, D., & Erşahin, M. (2019). A hybrid sentiment analysis method for Turkish. Turkish Journal of Electrical Engineering and Computer Sciences, 27(3), 1780-1793.
8. Gifari, M. K., Lhaksmana, K. M., & Dwifebri, P. M. (2021, October). Sentiment analysis on movie review using ensemble stacking model. In 2021 International conference advancement in data science, e-learning and information systems (ICADEIS) (pp. 1-5). IEEE.

9. Guven, Z. A. (2021, September). Comparison of BERT models and machine learning methods for sentiment analysis on Turkish tweets. In 2021 6th international conference on computer science and engineering (UBMK) (pp. 98-101). IEEE.
10. Habberrih, A., & Ali Abuzaraida, M. (2024, July). Sentiment analysis of Libyan dialect using machine learning with stemming and stop-words removal. In 5TH INTERNATIONAL CONFERENCE ON COMMUNICATION ENGINEERING AND COMPUTER SCIENCE (CIC-COCOS'24) (pp. 259-264). Cihan University-Erbil.
11. Iyer, V. (2024). A Comparative Analysis of Sentiment Classification Models for Improved Performance Optimization. Authorea Preprints.
12. Karagoz, P., Kama, B., Ozturk, M., Toroslu, I. H., & Canturk, D. (2019). A framework for aspect based sentiment analysis on turkish informal texts. Journal of Intelligent Information Systems, 53(3), 431-451.
13. Karah Bash, A. A., & Ercelebi, E. (2024). Advancing Sentiment Analysis during the Era of Data-Driven Exploration via the Implementation of Machine Learning Principles, Balkan Journal of Electrical and Computer Engineering, (2024), 12(1), 1-9.
14. Kılınç, D. (2019). A spark‐based big data analysis framework for real‐time sentiment prediction on streaming data. Software: Practice and Experience, 49(9), 1352-1364.
15. Kiziltepe, R. S., Ezin, E., & Karakus, M. (2024, October). A comparative study of ml models on large-scale turkish sentiment datasets. In 2024 Innovations in Intelligent Systems and Applications Conference (ASYU) (pp. 1-6). IEEE.
16. Mäntylä, M. V., Graziotin, D., & Kuutila, M. (2018). The evolution of sentiment analysis—A review of research topics, venues, and top cited papers. Computer science review, 27, 16-32.
17. Okereke, G. E., Udanor, C. N., Echezona, S. C., Asogwa, C. N., & Uzoh, I. (2024). Sentiment Analysis-An optimized Weighted Horizontal Ensemble approach. International Journal, 13(2).
18. Oliko, G. V., Muhambe, T. M., & Otieno, C. (2025). An Ensemble Model Based on Multinomial Naïve Bayes and Lexicon for Sentiment Classification of Product Reviews. International Journal of Computer Trends and Technology Volume 73 Issue 3, 1-15, March 2025.
19. Özçift, A., Akarsu, K., Yumuk, F., & Söylemez, C. (2021). Advancing natural language processing (NLP) applications of morphologically rich languages with bidirectional encoder representations from transformers (BERT): an empirical case study for Turkish. Automatika: časopis za automatiku, mjerenje, elektroniku, računarstvo i komunikacije, 62(2), 226-238.
20. Özmen, C. G., & Gündüz, S. (2025). Comparison of machine learning models for sentiment analysis of big Turkish web-based data. Applied Sciences, 15(5), 2297.
21. Rumelli, M., Akkuş, D., Kart, Ö., & Isik, Z. (2019, October). Sentiment analysis in Turkish text with machine learning algorithms. In 2019 Innovations in intelligent systems and applications conference (ASYU) (pp. 1-5). IEEE.
22. Sağbaş, E. A. (2024). A novel two-stage wrapper feature selection approach based on greedy search for text sentiment classification. Neurocomputing, 590, 127729.
23. Savci, P., & Das, B. (2023). Prediction of the customers' interests using sentiment analysis in e-commerce data for comparison of Arabic, English, and Turkish languages. Journal of King Saud University-Computer and Information Sciences, 35(3), 227-237.
24. Subba, B., & Kumari, S. (2022). A heterogeneous stacking ensemble based sentiment analysis framework using multiple word embeddings. Computational Intelligence, 38(2), 530-559.
25. Ucan, A., Naderalvojoud, B., Sezer, E. A., & Sever, H. (2016, November). SentiWordNet for new language: Automatic translation approach. In 2016 12th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS) (pp. 308-315). IEEE.
26. Wijati, D., Atika, P. D., Setiawati, S., & Rasim, R. (2024). Sentiment analysis of application reviews using the K-Nearest Neighbors (KNN) algorithm. PIKSEL: Penelitian Ilmu Komputer Sistem Embedded and Logic, 12(1), 209-218.
27. Yıldırım, M., Okay, F. Y., & Özdemir, S. (2020, December). Sentiment analysis for Turkish unstructured data by machine translation. In 2020 IEEE International Conference on Big Data (Big Data) (pp. 4811-4817). IEEE.
28. Zorarpaci, E. (2024, September). A Comparative Study of Ensemble Learning Methods for Turkish Text Categorization. In 2024 8th International Artificial Intelligence and Data Processing Symposium (IDAP) (pp. 1-6). IEEE.