A Framework for Applying Artificial Intelligence Components in the Information Retrieval Systems of Scientific Databases: A Meta-Synthesis and Fuzzy Delphi Study

Articles in Press

Document Type : Research َ Article

Authors

1 Department of Knowledge and Information Science, NT.C., Islamic Azad University, Tehran, Iran

2 Department of Computer Engineering, Shahab Danesh University, Qom, Iran

3 Department of Information and Knowledge Retrieval, Arak.C., Islamic Azad University

Abstract

Purpose: Traditional information retrieval systems in specialized journal databases struggle to meet complex user needs amidst the information explosion. Artificial Intelligence (AI), through natural language processing and personalization, offers significant potential for improvement. However, a comprehensive, systematic framework that identifies, evaluates, and prioritizes practical AI components for these specific systems is severely lacking. This study aims to fill this research gap by presenting a validated operational framework for the intelligent enhancement of these specialized database retrieval systems.

Method: This applied research used an exploratory mixed-methods approach. In phase one, a meta-synthesis of 112 selected articles (from 219 initial articles, 2000-2025) was conducted to extract 48 effective AI components. In phase two, a two-round Fuzzy Delphi technique with 19 experts was employed to validate and prioritize these components. Validity was confirmed by content validity, and reliability was confirmed by Cronbach's alpha (0.823).

Findings: Following the two-round Fuzzy Delphi evaluation, 33 components were ultimately confirmed as effective tools. These components were classified into seven main categories, ranked by expert consensus in order of importance: (1) "Intelligent Search and Recommendation," (2) "Advanced Analytics and Smart Applications," (3) "Natural Language Processing," (4) "Text Evaluation, Enrichment, and Improvement," (5) "Semantic Analysis, Organization, and Information Extraction," (6) "Multimedia Generation and Processing," and (7) "Language Translation and Conversion."

Conclusion: This research presents a comprehensive, prioritized framework that serves as a strategic roadmap for making specialized databases intelligent. The expert prioritization indicates a paradigm shift from a "repository of documents" to an "intelligent research assistant." The high ranking of "Intelligent Search and Recommendation" and "Advanced Analytics" highlights the critical importance of user experience and scientific integrity. This framework helps managers guide investments effectively.

Keywords

Main Subjects

Send comment about this article
CAPTCHA Image
Librarianship and Information Organization Studies

Articles in Press, Accepted Manuscript
Available Online from 29 April 2026

Files

History

  • Receive Date: 27 August 2025
  • Revise Date: 04 April 2026
  • Accept Date: 29 April 2026
  • Publish Date: 29 April 2026

Share

How to cite

Statistics