Browse Prior Art Database

A System and Method for efficient Services Asset Retrieval using iterative queries Disclosure Number: IPCOM000200896D
Publication Date: 2010-Oct-29
Document File: 3 page(s) / 128K

Publishing Venue

The Prior Art Database


We present a method for efficient information retrieval in iterative information seeking environment where multiple queries are used in an iterative way to locate the needed information. Each search query is a little refinement over the previous by changing or adding similar search terms to the existing query. Results of these multiple queries often overlap and users have to manually sift through them to get the wanted information. Existing approaches consider each refined query as independent of the previous queries which is typically not the case and hence the convergence of the search space becomes gradual and slow. A search query and its subsequent refined queries are not independent as the user repeatedly refines them till the desired information is found and we denote these queries to be a part of a single user session. This correlation of iterative queries is not reflected in the existing approaches.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 01 of 3

A System and Method for efficient Services Asset Retrieval using iterative queries


When searching for information on the internet, it is typically found that several queries have to be entered to locate the needed information. Each search query is a refinement over the previous by changing or adding a different set of search terms to get better search results. The user browses the search result and if not satisfied with the result or desires to see more information, refines the query and searches again. This is an iterative process and continues till the user seeks the desired information.

This process can be cumbersome as users typically do not know the right keywords for retrieving the desired information; they learn about the subject as the search proceeds and refine their keywords. It results in a gradual refinement of the search - for example, a user searching for 'self-managing systems' might learn during the course of the search process that self management of systems comes under the realm of autonomic computing and refines the search query to 'autonomic computing systems' or 'autonomic systems'.

Commercial search engines like AltaVista [1] and academic tools [2] [3] aid the user in refining the search query by recommending query modifications to reflect the user's information need more accurately thus narrowing down the search space. The big disadvantage of this process is that it is discontinuous. When a user selects a system recommended query, the previous search queries (used for retrieving the same information) are lost. Each of these multiple queries is treated as an independent query and the search results corresponding to the current query are returned. Discontinuity leads to an inferior user experience.

We describe an efficient information retrieval mechanism that retains the memory of queries until a 'session' is terminated. Session is ended when the user wants to search for something not related to the

previous query.


Two critical things to be addressed for this kind of search mechanism:
1. How to identify the session end during contiguous queries
2. How to maintain the coherence within a session

1. Determining session end

Problem statement: Given 2 sets of keyword search queries and their search results, determine if the 2nd keyword search is a new session or continuation of the previous one. In a continuous query system it is important to determine the relatedness of two search terms.
• If two search terms (set of key words) are related then probably the query is a variation of the

preceding queries.

• If the search terms are not related then possibly the quer...