180 likes | 379 Views
Information Retrieval Transfer Cycle. Dania Bilal IS 530 Fall 2006. Information Retrieval System. A set of components that interact to provide feedback Comprised of interlinked entities Agency that creates the databases People Documents. Interlinked Entities. Agency. People. Documents.
E N D
Information Retrieval Transfer Cycle Dania Bilal IS 530 Fall 2006
Information Retrieval System • A set of components that interact to provide feedback • Comprised of interlinked entities • Agency that creates the databases • People • Documents
Interlinked Entities Agency People Documents
IR Information Transfer • Inputs • Processes • Objectives of the System • Outputs
The IR Cycle • Documents are analyzed, translated, indexed, and stored. • Documents are organized • Cataloging (description/representation of docs.) • Subject indexing
The IR Cycle • Subject indexing a) Determination of subject content (conceptual analysis) b) Translation of content into language of the system (controlled vocabulary) c) Abstracting
The IR Cycle • Language of the system (controlled vocabulary) • List of subject headings (Pre-coordinate) • Thesauri (Pre-coordinate) • Classification scheme
The IR Cycle • Documents are represented by other entities • Author(s) • Date of publication • Language • Identifiers • Entities may become access points
The IR Cycle • Documents are stored after indexing • Document representation is entered into the matching mechanism • A file of document surrogates is established • File becomes available for searching using a variety of entities/access points
The IR Cycle • User Query • Analyzed for conceptual content • Translated into the language of the system (matched against controlled vocabulary and keywords) • Matched against document surrogates in the database
The IR Cycle • Output • A set of records found and deemed relevant to a user query • User judgment of retrieval
User Judgment • Relevance to information need • Relevance ranking by IR system • Relevance vs. pertinence
Document-Based IRs • Input, output, and matching mechanisms • Selection of documents (done by indexers) • Analysis of documents (done by indexers) • Document organization and representation • Done by indexers
Document-Based IRs • Analysis of user query (done by system) • Match of user query with relevant documents • Delivery of documents (output)
Information Seeking • Process of finding information to fill a knowledge gap • User requests • Known item searches • Unknown item searches Subject searches
Dialog System • Commercial vendor • Search available for a cost • Multi-subject system • Command- and menu-driven system
Dialog System • Four types of databases • Bibliographic (i.e., citations with or without abstracts) • Full-text (i.e., whole article) • Directory (e.g., Who’s Who) • Numeric (e.g., census, demographic)