QueryCat: automatic categorization of MEDLINE queries.

Area of Science:

• Information Science
• Computer Science
• Medical Informatics

Background:

• Information retrieval systems often yield overwhelming results due to suboptimal user queries.
• Effective query formulation is crucial for efficient access to relevant information.
• Organizing retrieved documents requires understanding the user's information need.

Purpose of the Study:

• To develop and evaluate a system for categorizing user queries.
• To improve the reformulation and organization of search results.
• To bridge the gap between user query and information need.

Main Methods:

• A two-phased approach for query categorization was implemented.
• Phase 1: Lexical analysis to identify query components.
• Phase 2: Semantic analysis to understand query meaning and intent.

Main Results:

• The developed system effectively categorizes user queries.
• Query categorization accuracy shows reasonable correspondence with expert assessments.
• The system's performance was validated against medical librarians and physicians.

Conclusions:

• Query categorization is a viable strategy to enhance information retrieval.
• The proposed system demonstrates effectiveness in abstracting user queries into common types.
• This approach facilitates improved query reformulation and document organization.