CURIOUS BROWSERS:
Automated Gathering of Implicit Interest Indicators
by an Instrumented Browser
David Brown and Mark Claypool
Computer Science Department
WPI, Worcester, MA 01609, USA
dcb@cs.wpi.edu claypool@cs.wpi.edu
ABSTRACT
Recommender systems provide personalized suggestions about items that users
will find interesting. Typically, recommender systems require a user interface that
can ``intelligently'' determine the interest of a user and use this information to
make suggestions. The common solution, ``explicit ratings'', where users tell the
system what they think about a piece of information, is well-understood and fairly
precise. However, having to stop to enter explicit ratings can alter normal
patterns of browsing and reading.
A more ``intelligent'' method is to use implicit ratings, where a rating is obtained
by a method other than obtaining it directly from the user. These implicit interest
indicators have obvious advantages, including removing the cost of the user
rating, and that every user interaction with the system can contribute to an implicit
rating.
Unfortunately, the ability of implicit ratings to predict actual user interest is not yet
well-understood. Our research studies the correlation between various implicit
ratings and the explicit rating for single Web pages.
We developed a a Web browser, called the Curious Browser, to record various
user actions (implicit ratings) while browsing and solicit explicit ratings for each
page. The Curious Browser captures mouse clicks, mouse movements, scrolling
and elapsed time.
In our first experiment, over 80 people used the Curious Browser to browse more
than 2500 Web pages [Le & Waseda 2000] [Claypool et al 2001]. Using the data
collected by the browser, the individual implicit ratings and some combinations of
implicit ratings were analyzed and compared with the explicit rating.
We found that the time spent on a page, the amount of scrolling on a page and
the combination of time and scrolling had a strong correlation with explicit
interest, while individual scrolling methods and mouse-clicks were ineffective in
predicting explicit interest.
In the second experiment [Law, Goodwin & Cen 2002], we collected 11 additional
indicators, including mouse wheel activity, status bar changes, the size of the
HTML file, the user's familiarity with the page, and a trace of the mouse
coordinates. Another set of around 80 users browsed nearly 1000 Web pages
using the Curious Browser. From the trace we extracted vertical and horizontal
lines, as our preliminary observations of users found some cursor following
activity over text and links that were of interest (See [Mueller and Lockerd,
2001]). We find some evidence that horizontal lines correlate with explicit
interest, and, not surprisingly, more total lines indicates more interest. Other
indicators did not appear to be significant under these circumstances.
Despite some difficulties in drawing strong conclusions from the second
experiment, we believe that mouse movement should be part of a set of simple
and compound implicit indicators that can be used to suggest a user's interest in
the content of a web page. Further studies need to be carried out to ascertain
stronger results.
References:
Mark Claypool, Phong Le, Makoto Waseda and David Brown. "Implicit Interest
Indicators", In Proceedings of ACM Intelligent User Interfaces Conference (IUI),
Santa Fe, New Mexico, USA, January 14-17, 2001.
Steve Law, Brad Goodwin & Michael Cen. "Curiouser Browsers", Undergraduate
Major Qualifying Project MQP-DCB-0106, Spring 2002, co-advised by Mark
Claypool and Dave Brown, Computer Science Department, WPI, Worcester, MA,
USA.
Phong Le and Makoto Waseda. "Curious Browsers", Undergraduate Major
Qualifying Project MQP-DCB-9906, Spring 2000, co-advised by Mark Claypool
and David Brown, Computer Science Department, WPI, Worcester, MA, USA.
Florian Mueller and Andrea Lockerd. "Cheese: Tracking Mouse Movements on
Websites, A Tool for User Modeling", Proceedings of ACM CHI Conference,
2001.
_________________________________
Version: 29th May 2003