Bibliometric information retrieval system (BIRS): A web search interface utilizing bibliometric research results

Ying Ding, Gobinda G. Chowdhury, Schubert Foo, Weizhong QianJ. Am. Soc. Inf. Sci 2000 · DOI

TL;DR

This paper introduces the Bibliometric Information Retrieval System (BIRS), a novel web-based interface designed to enhance information retrieval by integrating bibliometric analysis. BIRS aims to help users better understand research domains, formulate more effective queries, and discover relevant information by leveraging visualizations of author, journal, and keyword co-citation networks.

What Problem Does This Paper Address?

The authors tackle the persistent challenges in information retrieval (IR) stemming from information overload and the "vocabulary mismatch" problem. Traditional IR systems often struggle to help users effectively formulate and expand their search queries, especially when users are unfamiliar with the specific terminology of a domain. This leads to inefficient searches and a failure to retrieve relevant documents. Existing solutions like thesauri have limitations, and the sheer volume of information on the web makes it difficult for users to navigate and understand complex research landscapes.

Methods

The core of BIRS is its integration of bibliometric techniques, specifically:

  • Author Cocitation Analysis (ACA): Identifies prominent researchers and their intellectual connections by analyzing how often authors are cited together.
  • Journal Cocitation Analysis: Maps the relationships between journals based on their citation patterns, revealing interdisciplinary connections and core publications within a field.
  • Co-word Analysis: Analyzes the co-occurrence of keywords to identify important research topics and their interrelations, tracing the evolution of themes within a domain.

These bibliometric results are then visualized as "bibliometric maps" (cocitation maps) and integrated into a user-friendly web interface. BIRS also features:

  • Information Visualization: Presents the bibliometric maps to help users grasp the structure of a research domain.
  • Multilevel Browsing: Allows users to explore these maps at different levels of detail, from an overview to specific clusters and keywords.
  • Common User Interface: Provides a unified interface to search various sources, including web search engines, library OPACs, and online databases.

Datasets

The BIRS system is built upon the results of an ongoing bibliometric research effort focused on the Information Retrieval (IR) field. This research collected data from the Science Citation Index (SCI) and Social Science Citation Index (SSCI) for the period 1987-1997, encompassing 3,325 source papers from 971 journals with 78,785 citations.

Key Insights and Ideas

  • Bibliometrics as a Navigation Tool: The paper strongly advocates for using bibliometric analysis, particularly visualized maps, as a powerful tool to help users navigate complex research domains and understand their intellectual structures.
  • Bridging the Vocabulary Gap: By presenting authors, journals, and keywords that are highly related through cocitation or co-occurrence, BIRS directly addresses the vocabulary mismatch problem, offering users more relevant terms to refine their searches.
  • Integrated Search Experience: The concept of a "common user interface" that can interact with diverse search environments (web, OPAC, databases) is highlighted as crucial for a seamless and efficient user experience.
  • User-Centric Design: The system is designed with the end-user in mind, aiming to assist both novice and expert users in forming better queries and gaining deeper domain understanding.

Key Results

A preliminary user evaluation with 35 postgraduate students yielded positive feedback:

  • Improved Understanding: 80% of subjects reported a good or very good understanding of the IR area after using BIRS, with the keyword aspect being particularly helpful.
  • Query Formulation & Expansion: 77% agreed that BIRS significantly helps in forming and expanding queries. Specifically, 91% found the keyword aspect most useful, 34% found the author aspect helpful, and 29% found the journal aspect helpful.
  • Information Visualization: 80% gave good or very good comments on the information visualization feature, appreciating the ability to quickly identify interesting topics and zoom in on specific areas.
  • Common Search Interface: 69% found the interface easy to follow, and many noted its convenience in not needing to learn multiple search engines or databases.
  • Query Improvement: Subjects added new authors (43%), journals (26%), and keywords (63%) to refine their queries after using BIRS, indicating successful query expansion.

Limitations and Future Work

The authors acknowledge several limitations and suggest future directions:

  • System Issues: Some users found the maps static and not real-time, and experienced difficulty navigating or understanding their position within the maps.
  • Boolean Operator Sensitivity: The effectiveness of expanded queries was noted to be sensitive to the choice of Boolean operators (AND, OR, NOT), requiring careful user selection.
  • Extensibility: BIRS is designed to be extensible, with potential for incorporating more subject domains, additional map types, and new search engines to become a more comprehensive "one-stop tool" for information retrieval.
  • User Study Feedback: The user study provided valuable suggestions for improvement, which the authors intend to incorporate in future development.

Summary updated

BibTeX
@article{ding2000bibliometric,
  author = {Ying Ding and Gobinda G. Chowdhury and Schubert Foo and Weizhong Qian},
  title = {{Bibliometric information retrieval system (BIRS): A web search interface utilizing bibliometric research results}},
  journal = {J. Am. Soc. Inf. Sci.},
  year = {2000},
  doi = {10.1002/1097-4571(2000)9999:9999<::aid-asi1031>3.0.co;2-b},
}
Figure