Browse Prior Art Database

Intelligent Searching of Dynamic Databases Disclosure Number: IPCOM000122947D
Original Publication Date: 1998-Jan-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 4 page(s) / 156K

Publishing Venue


Related People

Wooding, GC: AUTHOR [+1]


The present disclosure relates to a data retrieval system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 38% of the total text.

Intelligent Searching of Dynamic Databases

      The present disclosure relates to a data retrieval system.

      A dynamic database is a database whose records are frequently
changed by its users.  Because the database is constantly updated,
the maintenance of a separate keyword index is complex.  A common
resolution to the problem of maintaining such an index, is for a
system to automatically use every word in a record as keywords to
that record.  When a search string is input by a user, search results
are based on the number of 'hits' per record, ie the number of
matches between the search string and keywords for a record.  This is
acceptable as long the search string and record keywords are
compatible; that is,  the required record contains at least one
occurrence of the entered search string.  This is not always possible
when the record includes references to an entity that has many
different names or ways of expressing that name - particularly
abbreviations, acronyms or nicknames.  Company names are a good
example of this: International Business Machines, IBM*, IBM Big Blue,

      In conventional data retrieval systems, the problem is passed
back to the user, who must amend the search string until the desired
record is hit.  Thus, the user's behavior must be adapted to the
content of the database.

      The present disclosure provides an intelligent link between the
search string and a record that, once established, can be re-used by
other users.  Thus, once I.B.M has been related to International
Business Machines, that linkage is recorded to save the frustration
of future users.

      The disclosure will now be described with reference to the
accompanying drawings, in which:

      Fig. 1 is a block diagram of a data structure employed in a
first embodiment, and Fig. 2 is a block diagram of a data structure
employed in a second embodiment.

      This disclosure allows a data retrieval system to adapt the way
it searches a large database based upon search strings preferred by
its users.

      This disclosure uses the concept of a keyword table that
contains actual search strings as supplied by users, together with
pointers into a master table that contains the data which is being
searched (Fig. 1).

      The concept is as follows, a user enters a search string,
according to the user's preference.  If the entered search string
exactly matches an existing keyword entry in the keyword table, the
index to the required entry in the master table can be returned
immediately to the system.  If there is no clear match between the
search string and a keyword table entry, then a list of those master
entries corresponding to the best partial matches in both the keyword
and master table is displayed.  The system prompts the user to make a
choice between these results.  Once the choice is made, the keyword
table is updated to associate the supplied search string with the
chosen master tab...