Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

A method of suggesting keywords for effective search

IP.com Disclosure Number: IPCOM000130490D
Original Publication Date: 2005-Oct-25
Included in the Prior Art Database: 2005-Oct-25
Document File: 2 page(s) / 34K

Publishing Venue

IBM

Abstract

Disclosed is a method for improving free-form text search efficiency by suggesting search keywords based on statistical analysis of keyword history.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 2

A method of suggesting keywords for effective search

This method helps search users to find relevant information easier and more quickly. Search users are often unfamiliar with special terms contained in documents. A number of search result documents are influenced by user's vocabularies when selecting a search keyword. This method improves free-form text search efficiency by suggesting search keywords.

These are the important characteristics of the method.

・ Showing keywords combined with three kinds of search, 'Include Search (AND)', 'OR Search', 'Not Include Search' ・ Automatic construction of database storing recommendation keywords. ・ Automatic maintenance of keywords according with how often the term is used, when the term is used the most recently

The system implementing this method consists of two components, Client (Browser) and Search Server. (Fig.1)

The all of functional components are placed into Search Server, and the entire data required for this mechanism is also placed into the Server. (Index and Search Data can be placed to other Server.)

The procedure of the mechanism is shown below.

I. Logging search activity (Search Pattern Logging Component)

1. Search user inputs some keywords with search conditions (AND, OR, NOT), then, executing queries.
2. Queries are sent to the Search Server.
3. Search Server shows the list of documents hit by the keywords.
4. The User picks up some of documents and refer to the content or sometimes don't.
5. Reference request is also sent to the Server.
6. The User often repeats this scenario through 1 to 5.
7. Server logs search activity (2, 5) into Search Pattern History Log File with information for identifying the user is the same person (ex. Client IP Address, UserID) and search date. The number of results is also logged with query sentence. And whether the document is referred by user is also logged. It is used for weight value.

II. Construction of Recommendation Keywords Database (Dictionary Construction Component) ・Construction is done periodically,...