August 16 2007–
\nThis summer I was in the Bay Area. I took the occasion and visit IBM Almaden Research and Google.
\nI gave a presentation both at IBM and Google on my research project, Gugubarra.<\/p>\n
Here are some details:<\/p>\n
Title of the talk:
\nThe Gugubarra Project: Building and Evaluating User Profiles for
\nVisitors of Web Sites.<\/p>\n
Speaker: Roberto V. Zicari
\nComputer Science Department,
\nGoethe University of Frankfurt, Germany.<\/p>\n
Work done in cooperation with Natascha Hoebel, Sascha Kaufmann,
\nKarsten Tolle, Naveed Mustaq.<\/p>\n
Index Terms: User Profiles, Web Communities, User modeling, Clustering methods<\/p>\n
—————————————————-
\nAbstract<\/p>\n
In this talk I will present an overview of the work we are
\ncurrently doing in the Gugubarra project. The project aims at
\nbuilding tools for better managing communities of Web visitors.<\/p>\n
The Gugubarra project (Gugubarra is the Aboriginal name for the
\nKookaburra bird) began in 2004 within the database group (DBIS)
\nat the Computer Science Institute of the Johann Wolfgang Goethe
\nUniversity, with the aim to build tools for better managing
\ncommunities of registered Web visitors.<\/p>\n
In this talk I will present the results of the project so far
\nand outline some open research issues.<\/p>\n
The starting point of our project is the assumption that a
\ncommunity of users is registered on a Web site and that for each
\nuser a profile is built. A User profile is based on the actions
\nand navigations the user performs on the Web site.<\/p>\n
In Gugubarra, we offer various settings that can be used to
\ncreate and manage user profiles. By using these settings the Web
\nsite owner can focus on those aspects he wants to analyze.<\/p>\n
The approach we have in Gugubarra is as follows:<\/p>\n
i) For each registered Web visitor we create a profile.
\nThese user profiles reflect the \u201cinferred\u201d interests of the
\nusers related to a set of pre-defined topics defined by the owner of
\nthe Web site. The profiles go beyond collecting the obvious
\ninformation the user is willing to give at the time of
\nregistration. In Gugubarra, a user profile contains two parts:
\nthe obvious profile, given directly by the user and a non
\nobvious profile (NOP), inferred by the user’s behavior during his visits
\non the site.<\/p>\n
ii) A user profile is (re)-calculated dynamically any time an
\nexplicit feedback is given by the user and\/or a set of events
\noccurred which are related to the user’s behavior and to certain
\n\u201clocations\u201d of the Web site.<\/p>\n
iii) We cluster Web visitors by clustering similar profiles of
\ninterest [5]. Cluster of Web visitors can then be used to
\nanalyze patterns of interests in the Web community and to forecast
\nfurther behavior. Clusters might also provide useful information
\nto support the decision what kind of new E-services to introduce
\nfor the Web community and when to introduce them.<\/p>\n
A first research prototype system, called Gugubarra 1.0 has been
\nimplemented in 2004 [1], which allows to build and manipulate
\nnon-obvious user profiles. It was showcased at the CeBIT Trade
\nFairs in 2004 and 2005. Gugubarra 1.0 works as a test
\napplication on real data provided by the Web community viewzone.org.<\/p>\n
A new prototype system, Gugubarra 2.0, is currently being
\ndesigned [3], [2] which includes a more sophisticated approach
\nto the definition of non-obvious user profiles and allows
\nclustering users by interests [4]. In this talk I will focus mainly on
\nthe new features introduced in Gugubarra 2.0 and refer to [6] for
\nthe features implemented in Gugubarra 1.0
\n—————————————————–<\/p>\n
Resources<\/p>\n
[1] Building and Evaluating Non-Obvious User Profiles for
\nVisitors of Web Sites. N. Mushtaq, K. Tolle, P. Werner and R.
\nV. Zicari.
\nIEEE Conference on E-Commerce Technology (CEC 04) July 6-9,
\n2004,San Diego, California, USA (.pdf 265KB)<\/p>\n
[2] The Gugubarra Project: Building and Evaluating User Profiles
\nfor Visitors of Web Sites. N. Hoebel, S. Kaufmann, K. Tolle, R.
\nV. Zicari.
\nFirst IEEE Workshop on Hot Topics in Web Systems and
\nTechnologies (HotWeb 2006), November 13-14, 2006, Boston,
\nMassachusetts, USA (.pdf 415)<\/p>\n
[3] The Design of Gugubarra 2.0: A Tool for Building and
\nManaging Profiles of Web Users. N. Hoebel, S. Kaufmann, K. Tolle, R. V.Zicari.
\nIEEE\/WIC\/ACM, International Conference on Web Intelligence,
\n2006, Hong Kong (.pdf 100) Long Version: (.pdf 196)<\/p>\n
[4] On Clustering Visitors of a Web Site by Behavior and
\nInterests.
\nN. Hoebel, R. V. Zicari. in Studies in Computational
\nIntelligence Series, Springer, AWIC ’07 (.pdf 155)<\/p>\n
The above papers can be downloaded at:
\nGugubarra Papers<\/a><\/p>\n