Search Assistance Research: Refining Search Patterns

June 7th, 2009 by Dr. Taly Weiss

Filed under information, internet behavior, Predictions, research, search, Trend Spotting Behavior, Trend Spotting Tools, Trends Review

TrendsSpotting has recently discussed the trend in which over the years people have extended the number of words used per search query. We have also addressed the existing gap between user’s search needs and search results.

A Penn State researchers analyzed and followed the search sessions of nearly 1 million Web searches to detect patterns of query reformulation and create models to predict them (and by that help create more advanced search engines).

Key findings:

The search terms in 22 percent of queries were reformulated or changed to more precisely convey the information for which the user was searching.
Users typically moved to narrow their query at the start of the session, moving to reformulation in the mid and latter portions of the sessions.
The assistance to narrow the query and alternate query terms would be most beneficial immediately after the initial query submission

According to the researchers “Given that one can predict future states of query formulation based on previous and present states with a reasonable degree of accuracy, one can design information systems that provide query reformulation assistance, automated searching assistance systems, recommender systems and others”.

(Jim Jansen co-authored the research paper with Danielle Booth, Penn State information sciences and technology student and Amanda Spink, Queensland University of Technology, Australia).

With this week launch of Bing, when many of us tested and compared search success – it is obvious that a smart recommendation system can be much help.

We will be waiting!

Tweet This Post Plurk This Post Buzz This Post Delicious Digg This Post Ping This Post Reddit Stumble This Post