Browse Prior Art Database

Agent System for Gathering, Integrating, Relevance Ranking and Presenting Digital Text Documents from Heterogeneous Information Sources

IP.com Disclosure Number: IPCOM000122844D
Original Publication Date: 1998-Jan-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 2 page(s) / 92K

Publishing Venue

IBM

Related People

Callaghan, PJ: AUTHOR [+4]

Abstract

Disclosed is an agent system that allows multiple heterogeneous information sources of digital text documents to be searched in response to a user's query. A user enters a single query into the agent system. The agent system submits that query against the different information sources and creates a temporary database containing all the documents from all the information sources that satisfy the query. The documents in the temporary database are ranked according to how relevant they are to the user's query and are presented to the user in a single ordered list.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Agent System for Gathering, Integrating, Relevance Ranking and Presenting
Digital Text Documents from Heterogeneous Information Sources

      Disclosed is an agent system that allows multiple heterogeneous
information sources of digital text documents to be searched in
response to a user's query.  A user enters a single query into the
agent system.  The agent system submits that query against the
different information sources and creates a temporary database
containing all the documents from all the information sources that
satisfy the query.  The documents in the temporary database are
ranked according to how relevant they are to the user's query and are
presented to the user in a single ordered list.

      The user interface to the agent system is via a web
browser.  To use the system, a user enters a query (#1 in the
Figure).  A query consists of two separate pieces of data: a list of
keywords and a  free-format text description of the type of
information that the user will like to see.

      A Boolean query is formed by creating the conjunction of each
keyword that the user entered.  The agent system sends that Boolean
query to each information source, shown as #2 in the Figure.  Each
information source performs the Boolean query against the documents
in its repository and returns the matching documents in an ASCII
format.  When an information source has completed returning all
matching documents, it tells the agent system that it is done with
the query.  All communication between the information sources and the
agent system is through an adapter that is responsible for providing
the interface to the information source.

      As matching documents are returned to the agent system, they
are added to a temporary database (#3 in the Figure) created by the
agent system.  A separate database is created for each query entered
by users.  When all information sources have indicated that they have
completed the Boo...