Browse Prior Art Database

Automatic Search Type Detection

IP.com Disclosure Number: IPCOM000126798D
Original Publication Date: 2005-Aug-02
Included in the Prior Art Database: 2005-Aug-02
Document File: 1 page(s) / 39K

Publishing Venue

IBM

Abstract

A common problem with searching is that users have to manually narrow the scope of the search. If a user does not narrow the scope, then everything is searched or the search is rejected. If everything is searched then it is costly in performance and also provides less useful search results. If the search is rejected then the user has to start again. This article describes a method for automatically narrowing the scope of a search.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 66% of the total text.

Page 1 of 1

Automatic Search Type Detection

This paper describes a searching method that automatically narrows the scope of a search. Currently a user has to manually narrow the search, usually by choosing a search type. This adds an extra step into the process of searching. If a user does not select a type, then everything will be searched. Some search systems will not even run the search unless a type is selected. The automation is achieved by working out what type of search should be carried out, based on the search term enetered by the user. This improves performance and also the quality of search results.

    As an example, there is a search facility on an employee directory. There are many types of data you can search on, for example, a person's name, phone number, email address, etc. Currently a user would select which of these types you wish to search on, then type in your search term. To automate this narrowing of scope, the type must be determined from the search term, so for example, if a phone number was entered then only phone numbers would be searched. This avoids the performance issues of searching all types, and removes the extra manual step.

    To achieve this automation, a regular expression is defined for each search type. A search term entered by a user is compared against these expressions. When a match is found, the search type is known. So for example, we could define the email address regular expression as:

[ a-z , A-Z , 0-9 , _ , . ]+ @ [ a-z , A-Z , 0-9 ,...