SavedAdd your journal

Journals / Beykent Üniversitesi Fen ve Mühendislik Bilimleri Dergisi / 2026 / Cilt: 19 Sayı: 1

A Hybrid Artificial Intelligence-Supported Approach for Semantic Category Matching from Free-Form Text Inputs

Serbest Metin Girdilerinden Anlamsal Kategori Eşleştirmesi İçin Hibrit Yapay Zekâ Destekli Bir Yaklaşım

Ezgi Gökdemir(Sabancı Dijital Teknoloji Hizmetleri A.Ş.)0009-0004-1701-721X
Kemal Elçi(Sabancı Dijital Teknoloji Hizmetleri A.Ş.)0009-0003-3698-1867
Hazal Oğuz Azğüler(Sabancı Dijital Teknoloji Hizmetleri A.Ş.)0009-0008-3917-4602
Pages
55–65
Publication Date
2026-06-29
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

Studies conducted on a hierarchical structure with approximately 14,000 categories demonstrate that the proposed method is scalable and delivers highly accurate results.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:The correct categorization of products and services offered by suppliers on e-procurement platforms is critical for search and operational efficiency; however, matching traditional keyword search methods with multi-level category structures is time-consuming and produces inaccurate results under current conditions.
📊
Key Findings & Data:This study presents a hybrid AI-supported approach for automatic category matching suggestions from users' free-text inputs.
💡
Conclusion & Impact:Studies conducted on a hierarchical structure with approximately 14,000 categories demonstrate that the proposed method is scalable and delivers highly accurate results.
🔬Design:Araştırma Makalesi
📍Scope:Comprehensive Sample
⚙️Variables:Anlamsal Benzerlik · Doğal Dil İşleme · E-Satın Alma

Abstract

The correct categorization of products and services offered by suppliers on e-procurement platforms is critical for search and operational efficiency; however, matching traditional keyword search methods with multi-level category structures is time-consuming and produces inaccurate results under current conditions. This study presents a hybrid AI-supported approach for automatic category matching suggestions from users' free-text inputs. The method performs category suggestions by combining semantic similarity, orthographic similarity, and usage popularity. Studies conducted on a hierarchical structure with approximately 14,000 categories demonstrate that the proposed method is scalable and delivers highly accurate results.

Özet

E-satın alma platformlarında tedarikçilerin sundukları ürün ve hizmetler için doğru kategori eşleşmelerinin bulunması, arama ve operasyonel verimlilik açısından kritik öneme sahiptir; ancak geleneksel anahtar kelime aramasıyla çok seviyeli kategori yapıları arasındaki eşleştirme, mevcut koşullarda zaman alıcı ve düşük doğruluklu sonuçlar üretmektedir. Bu çalışmada, kullanıcıların serbest metin girdilerinden otomatik kategori eşleştirme önerileri sunmak için hibrit yapay zekâ destekli bir yaklaşım önerilmektedir. Yöntem; anlamsal benzerlik, yazımsal benzerlik ve kullanım popülerliğini ağırlıklı olarak birleştirerek kategori önerisi gerçekleştirmektedir. Yaklaşık 14.000 kategorili hiyerarşik yapı üzerinde yapılan çalışmalar, önerilen yöntemin ölçeklenebilir ve yüksek doğruluklu sonuçlar sunduğunu göstermektedir.

🔖

Cite This Article

Gökdemir, E., Elçi, K., Azğüler, H. O. (2026). A Hybrid Artificial Intelligence-Supported Approach for Semantic Category Matching from Free-Form Text Inputs. Beykent Üniversitesi Fen ve Mühendislik Bilimleri Dergisi, Cilt: 19 Sayı: 1, 55–65. https://doi.org/10.20854/bujse.1919769

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (14)

  1. Balli, N., & Türkoğlu Kaya, T. (2025). Kullanıcı eğilimlerinin göreceli modellenmesinin popülerlik yanlılığına etkisi. Çukurova Üniversitesi Mühendislik Fakültesi Dergisi, 40(3), 671–686. https://doi.org/10.21605/cukurovaumfd.1748924
  2. Cheng, X., Bowden, M., Bhange, B. R., Goyal, P., Packer, T., & Javed, F. (2021). An end-to-end solution for named entity recognition in eCommerce search. The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21), 15098–15106. https://doi.org/10.1609/aaai.v35i17.17773
  3. Cheng, Z., Zhang, W., Chou, C. C., Jau, Y. Y., Pathak, A., Gao, P., & Batur, U. (2024). E-commerce product categorization with LLM-based dual-expert classification paradigm. Proceedings of the 1st Workshop on Customizable NLP: Progress and Challenges in Customizing NLP for a Domain, Application, Group, or Individual (CustomNLP4U), 294–304. https://doi.org/10.18653/v1/2024.customnlp4u-1.22
  4. Çiftlikçi, M. S., Çakmak, Y., Kalaycı, T. A., Abut, F., Akay, M. F., & Kızıldağ, M. (2025). A new large language model for attribute extraction in e-commerce product categorization. Electronics, 14(10), 1930. https://doi.org/10.3390/electronics14101930
  5. Gholamian, S., Romani, G., Rudnikowicz, B., & Skylaki, S. (2024). LLM-based robust product classification in commerce and compliance. Proceedings of the 1st Workshop on Customizable NLP: Progress and Challenges in Customizing NLP for a Domain, Application, Group, or Individual (CustomNLP4U), 26–36. https://doi.org/10.18653/v1/2024.customnlp4u-1.3
  6. Hyppänen, T. (2025). AI possibilities in business process development: A case study in public procurement [Yüksek lisans tezi, Lappeenranta–Lahti University of Technology LUT].
  7. Karabacak, O., & Inan, E. (2025). A semantic search engine for Turkish and English research resources. 2025 Innovations in Intelligent Systems and Applications Conference (ASYU), 1–5. https://doi.org/10.1109/ASYU67174.2025.11208360
  8. Şirin, Y., & Kutlugün, M. A. (2020). Anlamlı ve benzer olmayan Türkçe metinler üretmek için N-Gram yöntemi ile istatistiksel ve kural tabanlı yaklaşımın birlikte kullanımı. Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi, 22(65), 331–342. https://doi.org/10.21205/deufmd.2020226504
  9. Jahanshahi, H., Ozyegen, O., Cevik, M., Bulut, B., Yigit, D., Gonen, F. F., & Başar, A. (2021). Text classification for predicting multi-level product categories (arXiv Preprint arXiv:2109.01084). arXiv. https://arxiv.org/abs/2109.01084
  10. Liu, J., & Li, P. (2026). Large language models as decision support tools in procurement and supplier selection. Frontiers in Artificial Intelligence Research, 3(1), 167–186. https://doi.org/10.71465/fair714
  11. Ronchi, S., Brun, A., Golini, R., & Fan, X. (2010). What is the value of an IT e-procurement system?. Journal of Purchasing and Supply Management, 16(2), 131–140. https://doi.org/10.1016/j.pursup.2010.03.013
  12. Schoenherr, T., & Tummala, V. M. R. (2007). Electronic procurement: A structured literature review and directions for future research. International Journal of Procurement Management, 1(1-2), 8–37. https://doi.org/10.1504/IJPM.2007.015353
  13. Steck, H., Ekanadham, C., & Kallus, N. (2024). Is cosine-similarity of embeddings really about similarity?. WWW '24: Companion Proceedings of the ACM Web Conference, 887–890. https://doi.org/10.1145/3589335.3651526 Stenström, A., & Cederberg, N. (2025). Scalable optimization of product category embeddings using multi-dimensional scaling and LLM embeddings [Yüksek lisans tezi, Lund University].
  14. Uysal, S., Baysan, M. S., İşlek, İ., & Karaman, Ç. Ç. (2025). Semantic search approach for classified platforms. 33rd Signal Processing and Communications Applications Conference (SIU), 1–4. https://doi.org/10.1109/SIU66497.2025.11112423