AI evidence extraction

Intelligent Semantic Search for Academic Journals Using AI and NLP Techniques

Shireen Fathi Malo · Journal of Information Systems Engineering & Management · 2025

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

Evidence

Citations

5.72

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.52783/jisem.v10i41s.7884

Methodology & findings

Study design

System design and implementation study.

Primary method

Design science research approach combining system design, architecture, and integration of existing AI/NLP technologies

Main result

The study demonstrates that "the proposed system demonstrates an effective approach to understanding and responding to complex natural language queries" through integration of semantic embeddings, linguistic preprocessing, and ontology-based expansion. Prior literature shows that "AI-powered search engines provide more relevant results, reduce user frustration, and transform research practices by enabling efficient and intuitive access to academic resources," and experiments on large-scale datasets "demonstrated significant improvements: a 15% increase in Normalized Discounted Cumulative Gain (NDCG) for complex queries and a 12% improvement in Mean Reciprocal Rank (MRR) for navigational queries, compared to traditional keyword-based approaches."

Research paradigm

Positivist/Empiricist

Author conclusions

"This study presents a comprehensive and scalable framework for semantic academic journal search by integrating advanced AI and NLP techniques into a hybrid retrieval system." The authors conclude that "unlike traditional keyword-based systems, this solution bridges the semantic gap between user intent and academic content, supporting more accurate, context-aware discovery of peer-reviewed journals" and that "the contributions of this research not only advance the field of intelligent information retrieval but also offer practical value for researchers, librarians, and academic institutions aiming to streamline access to high-impact scholarly resources."

Risk of bias

No empirical validation or user study reported - system evaluated theoretically only; No quantitative performance metrics or benchmark comparisons provided; Selection bias in data sources: data aggregated from Scopus and DOAJ only; No discussion of potential biases in the Sentence Transformer model (all-mpnet-base-v2); Ontology coverage not specified - potential bias in domain concept representation; reliance on theoretical framework without empirical testing; potential bias in journal selection (Scopus and DOAJ may not represent all academic journals); no comparison with other semantic search systems on same dataset; limited discussion of confounding factors in retrieval ranking.

Open questions raised

Future work may explore integration of cross-lingual capabilities, personalized search recommendations, and adaptive learning mechanisms to further optimize performance and user experience. Additional challenges identified in literature include factual accuracy and conflicting opinions in AI research navigation platforms.

Data: Journal metadata aggregated from Scopus and Directory of Open Access Journals (DOAJ); specific URLs for dataset access not provided in paper; benchmark dataset mentioned as contribution but not explicitly made available; SciTLDR dataset referenced in related workExtracted from: pdf

Explore related topics

AI-Assisted Discovery Literature Discovery Design Science