Metin Analitiği ve Regresyon Analizi Yoluyla İkinci El Pazar Uygulamalarındaki Genel Derecelendirmelerin Yön-Duygu Sürücülerinin Anlaşılması

Abstract

Kullanıcıların ürettiği yorumlar ve yıldız değerlendirmeleri, mobil uygulama pazarlarında müşteri güveni, indirme kararları ve platform itibarı üzerinde güçlü bir etkiye sahiptir. Bu çalışma, Google Play’deki üç Türk ikinci el pazar yeri uygulamasının (Letgo, Dolap, Gardrops) yorumlarını inceleyerek, belirli boyut–duygu örüntülerinin genel değerlendirme puanlarını nasıl etkilediğini ortaya koymaktadır. Çok aşamalı bir doğal dil işleme (NLP) süreci uygulanmıştır. Yorum cümleleri çok dilli SBERT ile vektörleştirilmiştir. KMeans yöntemiyle kümelenmiş ve 13 üst düzey boyut kategorisi elde edilmiştir. Duygu sınıflandırması ise, 1.000 elle etiketlenmiş cümle üzerinde doğrulanan, alan özelinde Türkçe ELECTRA modeli kullanılarak yapılmıştır. Ardından Ridge regresyonu ile boyut–duygu çiftlerinin yıldız puanlarına katkısı nicel olarak ölçülmüştür. Analiz sonuçları, iade, kargo maliyetleri ve sahtekârlıkla ilgili olumsuz deneyimlerin puanları düzenli olarak düşürdüğünü, buna karşılık kullanılabilirlik, işlem memnuniyeti ve müşteri desteğine ilişkin olumlu ifadelerin daha güçlü puan artışları sağladığını göstermiştir. Karşılaştırmalı bulgular, daha yüksek puanlı uygulamaların (Dolap: 4,4; Gardrops: 4,2) daha fazla olumlu deneyim biriktirdiğini, Letgo’nun (2,7) ise tekrar eden güven ve adalet sorunları sergilediğini ortaya koymuştur. Bu sonuçlar, müşteri değerlendirmelerini en güçlü şekilde şekillendiren hizmet boyutlarını vurgulamaktadır. Çalışma, boyut düzeyinde duygu ile puan sonuçlarını ilişkilendirerek, ikinci el pazar yerlerinde müşteri memnuniyetini artırma ve rekabet gücünü geliştirme konusunda platform yöneticilerine uygulanabilir içgörüler sunmaktadır.

Keywords

• Özellik Tabanlı Duygu Analizi
• Ridge Regresyonu
• Müşteriden Müşteriye Platformlar
• Metin Analitiği
• Çevrim İçi Derecelendirmeler

Understanding Aspect-Sentiment Drivers of Overall Ratings in Second Hand Marketplace Apps through Text Analytics and Regression Analysis

Abstract

User-generated reviews and star ratings strongly influence customer trust, download decisions, and platform reputation in mobile app marketplaces. This study analyzes reviews from three Turkish secondhand marketplace applications (Letgo, Dolap, Gardrops) on Google Play to uncover how specific aspect–sentiment patterns affect overall ratings.A multi-stage natural language processing (NLP) pipeline was applied. Review sentences were embedded with multilingual SBERT and clustered using KMeans, resulting in 13 higher level aspect categories. Sentiment was classified using a domain specific Turkish ELECTRA model validated on 1,000 manually annotated sentences. Ridge regression was then employed to quantify the contribution of aspect–sentiment pairs to star ratings. The analysis showed that negative experiences related to returns, shipping costs and fraudulent practices consistently decreased ratings, while positive mentions of usability, transaction satisfaction and customer support produced stronger rating improvements. Comparative findings revealed that higher-rated apps (Dolap: 4.4, Gardrops: 4.2) accumulated more positive experiences, whereas Letgo (2.7) exhibited recurring trust and fairness issues. These results highlight which service aspects most strongly shape customer evaluations. By linking aspect-level sentiment to rating outcomes, the study provides platform managers with actionable insights for improving satisfaction and strengthening competitiveness in secondhand C2C marketplaces.

Keywords

• Aspect-Based Sentiment Analysis
• Ridge Regression
• Customer-to-Customer (C2C) Platforms
• Text Analytics
• Online Ratings

References

1. [1] M. Harman, Y. Jia, and Y. Zhang, “App store mining and analysis: MSR for app stores,” in Proc. 9th IEEE Working Conf. Mining Softw. Repositories (MSR’12), 2012, pp. 108–. doi: 10.1145/2804345.2804346.
2. [2] H. Sällberg, S. Wang, and E. Numminen, “The combinatory role of online ratings and reviews in mobile app downloads: An empirical investigation of gaming and productivity apps from their initial app store launch,” J. Marketing Analytics, vol. 11, no. 3, pp. 426–442, 2022. doi: 10.1057/s41270-022-00171-w.
3. [3] S. Ba, S. He, and S. Lee, “Mobile app adoption and its differential impact on consumer shopping behavior,” Prod. Oper. Manag., vol. 31, no. 2, pp. 764–780, 2022. doi: 10.2139/ssrn.3727035.
4. [4] Y. Amirkhalili and H. Y. Wong, “Banking on feedback: Text analysis of mobile banking iOS and Google app reviews,” arXiv Preprint, arXiv:2503.11861, 2025. doi: 10.48550/arXiv.2503.11861.
5. [5] L. Zhang, S. Wang, and B. Liu, “Deep learning for sentiment analysis: A survey,” Wiley Interdiscip. Rev. Data Mining Knowl. Discov., vol. 10, no. 3, e1331, 2020. doi: 10.1002/widm.1253.
6. [6] S. Vanaja and M. Belwal, “Aspect-level sentiment analysis on e-commerce data,” in 2018 Int. Conf. Inventive Res. Comput. Appl. (ICIRCA), 2018, pp. 1275–1279. doi: 10.1109/ICIRCA.2018.8597286.
7. [7] Y. Wang, Y. Huang, and M. Wang, “Aspect-based rating prediction on reviews using sentiment strength analysis,” in Int. Conf. Ind., Eng. Other Appl. Appl. Intell. Syst., Springer, 2017, pp. 439–447. doi: 10.1007/978-3-319-60045-1_45.
8. [8] Z. Drus and H. Khalid, “Sentiment analysis in social media and its application: Systematic literature review,” Procedia Comput. Sci., vol. 161, pp. 707–714, 2019. doi: 10.1016/j.procs.2019.11.174.

9. [9] K. Aziz, D. Ji, P. Chakrabarti, T. Chakrabarti, M. S. Iqbal, and R. Abbasi, “Unifying aspect-based sentiment analysis BERT and multi-layered graph convolutional networks for comprehensive sentiment dissection,” Sci. Rep., vol. 14, no. 1, 14646, 2024. doi: 10.1038/s41598-024-61886-7.
10. [10] R. J. Dhanal and V. R. Ghorpade, “Aspect-based sentiment analysis using topic modelling and machine learning,” Int. J. Electr. Comput. Eng., vol. 14, no. 6, 2024. doi: 10.11591/ijece.v14i6.pp6689-6698.
11. [11] Y. C. Hua, P. Denny, J. Wicker, and K. Taskova, “A systematic review of aspect-based sentiment analysis: Domains, methods, and trends,” Artif. Intell. Rev., vol. 57, no. 11, 296, 2024. doi: 10.1007/s10462-024-10906-z.
12. [12] A. C. Öztürk, “Large Group Decision Making for Aspect-Level Consensus Evaluation in Low-Rated App Reviews”, Müh.Bil.ve Araş.Dergisi, c. 7, sy. 2, ss. 173–184, 2025, doi: 10.46387/bjesr.1716998.
13. [13] L. Davoodi, J. Mezei, and M. Heikkilä, “Aspect-based sentiment classification of user reviews to understand customer satisfaction of e-commerce platforms,” Electron. Commer. Res., pp. 1–43, 2025. doi: 10.1007/s10660-025-09948-4.
14. [14] S. Gojali and M. L. Khodra, “Aspect-based sentiment analysis for review rating prediction,” in 2016 Int. Conf. Adv. Informatics: Concepts, Theory Appl. (ICAICTA), 2016, pp. 1–6. doi: 10.1109/ICAICTA.2016.7803110.
15. [15] Y. Putranto, B. Sartono, and A. Djuraidah, "Topic modelling and hotel rating prediction based on customer review in Indonesia," International Journal of Management and Decision Making, vol. 20, no. 3, pp. 282–307, 2021.
16. [16] F. Hawlitschek, T. Teubner, and C. Weinhardt, “Trust in the sharing economy,” Die Unternehmung, vol. 70, no. 1, pp. 26–44, 2016. doi: 10.5771/0042-059X-2016-1-26.
17. [17] E. Ert, A. Fleischer, and N. Magen, "Trust and reputation in the sharing economy: The role of personal photos in Airbnb," Tour. Manag., vol. 55, pp. 62–73, 2016, doi: 10.2139/ssrn.2624181.
18. [18] D. Demirol, R. Das, and D. Hanbay, ‘A key review on security and privacy of big data: issues, challenges, and future research directions’, SIViP, Sept. 2022, doi: 10.1007/s11760-022-02341-w.
19. [19] D. Demirol, R. Das, and D. Hanbay, ‘Büyük veri üzerine perspektif bir bakış’, in 2019 International Artificial Intelligence and Data Processing Symposium (IDAP), Sept. 2019, pp. 1–9. doi: 10.1109/IDAP.2019.8875902.
20. [20] Akın, A. A., & Akın, M. D. (2007). Zemberek, an open source NLP framework for Turkic languages. In Proceedings of the 3rd International Balkan Conference on Communications and Networking (BalkanCom). Istanbul, Turkey.
21. [21] R. Güran, “Stopword list for Turkish,” GitHub, [Online]. Available: https://github.com/stopwords-iso/stopwords-tr.
22. [22] K. Clark, M.-T. Luong, Q. V. Le, and C. D. Manning, “ELECTRA: Pre-training text encoders as discriminators rather than generators,” in Int. Conf. Learn. Represent. (ICLR), 2020. doi: 10.48550/arXiv.2003.10555.
23. [23] M. Laurer, mDeBERTa-v3-base-xnli-multilingual-nli-2mil7 [Model]. Hugging Face, 2021. [Online]. Available: https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-xnli-multilingual-nli-2mil7.
24. [24] C. D. Manning, P. Raghavan, and H. Schütze, Introduction to Information Retrieval. Cambridge, U.K.: Cambridge Univ. Press, 2008. doi: 10.1017/CBO9780511809071.
25. [25] F. Nurifan, R. Sarno, and K. R. Sungkono, “Aspect based sentiment analysis for restaurant reviews using hybrid elmo-wikipedia and hybrid expanded opinion lexicon-senticircle,” International Journal of Intelligent Engineering and Systems, vol. 12, no. 6, pp. 47–58, 2019, doi: 10.22266/ijies2019.1231.05.
26. [26] G. James, D. Witten, T. Hastie, and R. Tibshirani, An Introduction to Statistical Learning with Applications in R, 2nd ed. New York, NY, USA: Springer, 2021. doi: 10.1007/978-1-0716-1418-1.