Web personalization using implicit input
Loading...
Date
2007
Authors
Pan, Yih Jyh
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The growing importance of the World Wide Web in our lives has intensified the
studies on personalization. These studies on personalization generally make use of
explicit information, e.g. rating an item to know the interests or disinterests of users.
However, this method of obtaining information is intrusive on the users. As a result,
users often shy away from updating their likes and dislikes. Consequently, their latest
interests are not known.
Hence, my work seeks to look for an alternative way to obtain input from users
in a less obtrusive manner, namely implicit input. From my studies, majority of the
researches of the use of implicit input are focusing on capturing positive interests of
users. However, the disinterests of users are often neglected.
This intrigues me to find out the possibility of using implicit input to capture the
disinterests of users as well. Hence, I categorize my selection of implicit input into two
groups: (a) positive interest indicators, viz. view, book-mark, add-to-cart, and purchase,
and (b) negative interest indicators, viz. skip, delete book-mark, delete from cart.
A simulated online shopping mall is used in my work to observe and gather
information from my users. I am able to come to a conclusion that implicit input is
indicative of user interests, but there is no clear support to show that implicit input can
be suitably used to reflect the disinterests of users.
In the final part of my methodology, I adopted a few strategies of inferring
feedback ratings from implicit input, that I embrace could be applied to replace explicit
user ratings. As a result, I demonstrate that the list of implicit input studied in my work
can be used to generate tangible output, which in turn can be helpful in predicting user
interests.
Description
Master
Keywords
Computer Science , Web personalization