Proceeding of the Faculty of technical Sciences, Novi Sad

Electrotechnical and Computer Engineering

Vol. 41 No. 10 (2026): Proceedings of the Faculty of Technical Sciences

Comparative Analysis of Selected Retrieval Optimization Approaches in RAG Architecture

  • Сања Петровић
DOI:
https://doi.org/10.24867/36BE52Petrovic
Submitted
October 14, 2026
Published
2026-10-16

Abstract

Pre-trained language models often lack knowledge of narrowly specialized domains such as pharmacy, and the use of the Serbian language poses an additional challenge. For this reason, a Retrieval-Augmented Generation (RAG) architecture is needed, which extends a generative large language model with an external knowledge base. There are numerous approaches for improving retrieval within a RAG architecture, but it is necessary to analyze and evaluate their impact in the context of the pharmacy domain in the Serbian language. Therefore, a comparative analysis experiment was conducted on the impact of metadata-based filtering, hybrid search, diversification, and query transformation approaches, the results of which are presented in this paper.

References

  1. [1]
  2. A. Bansal, “Optimizing RAG with Hybrid Search and Contextual Chunking,” Journal of Engineering and Applied Sciences Technology, vol. 5, 2023, doi: 10.47363/JEAST/2023(5)E114.
  3. [2]
  4. Y. Gao, Y. Xiong, X. Gao, K. Jia, J. Pan, Y. Bi, Y. Dai, J. Sun, M. Wang, and H. Wang, "Retrieval-augmented generation for large language models: A survey," arXiv preprint arXiv:2312.10997, 2023.
  5. [3]
  6. M. Poliakov and N. Shvai, “ Multi-Meta-RAG: Improving RAG for Multi-Hop Queries using Database Filtering with LLM-Extracted Metadata,” in International Conference on Information and Communication Technologies in Education, Research, and Industrial Applications, Cham: Springer Nature Switzerland, 2024, doi: 10.1007/978-3-031-81372-6_25.
  7. [4]
  8. J. G. Carbonell and J. Goldstein, “The Use of MMR, Diversity-Based Reranking for Reordering Documents and Producing Summaries,” in Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '98), 1998, doi: 10.1145/290941.291025.
  9. [5]
  10. S. Es, J. James, L. E. Anke, and S. Schockaert, “RAGAS: Automated Evaluation of Retrieval Augmented Generation“ in Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations, 2024, doi: 10.18653/v1/2024.eacl-demo.16.
  11. [6]
  12. https://docs.ragas.io/en/stable/concepts/metrics/available_metrics/ (приступљено у мају 2026.)
  13. [7]
  14. https://www.alims.gov.rs/humani-lekovi/pretrazivanje-humanih-lekova/ (приступљено у мају 2026.)