PAZZANI, M. and BILLSUS, D. 1997, Learning and revising user proles: the identication of interesting web sites. Machine Learning, 27, 313331. @ NUS Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Personalized Webdocument Filtering Using Reinforcement Learning - Zhang, Seo (2001) (1 citation) (Correct)
....AITRC, and BK21 IT Program. Address correspondence to Dr. Byoung Tak Zhang, Biointelligence Lab, School of Computer Science and Engineering, Seoul National University, Seoul 151 742, Korea. E mail: btzhang bi.snu.ac. kr 665 users during the interaction with the information retrieval systems (Pazzani Billsus, 1997). Several methods have been proposed to re ect user preferences (Kindo et al. 1997) A classical approach is the Rocchio method ( 1971) and its variants. This is a batch that modi es the original query vector by the vectors of the relevant and irrelevant documents. However, the batch algorithms ....
Pazzani, M., and D. Billsus. 1997. Learning and revising user pro les: the identi cation of interesting Web sites. Machine L earning 27:313 331.
Learning About the User: A General Approach and Its.. - Pohl, Schwab, Koychev (Correct)
....representation reasoning adaptive feature decision acquisition KR system information flow component acquisition user modeling task Figure 1: Using a knowledge representation system for user adapted interaction. ltering have been described in the literature, like Syskill Webert [ Pazzani and Billsus, 1997 ] Letizia [ Lieberman, 1995 ] or Amalthaea [ Moukas, 1996 ] In general, machine learning (ML) methods process training input and o er support for decision (mainly classi cation) problems based on this input. Hence, ML based user adaptive systems work quite di erently from KR based ones. ....
....solve classi cation problems. Hence, a straightforward way of using machine learning for acquiring interest pro les is to assume that the set of information objects can be divided into classes (e.g. interesting and not interesting ) and to provide examples for both classes (see, e.g. Pazzani and Billsus, 1997 ] Also in ELFI it can be assumed that there are two such classes of DVs for each user. However, supplying an appropriate set of negative examples is problematic. There are systems that use unselected or un viewed objects as negative examples. In ELFI at least, unselected DVs may exist that are ....
[Article contains additional citation context not shown here]
M. Pazzani and D. Billsus. Learning and revising user proles: The identi cation of interesting web sites. Machine Learning, 27:313-331, 1997.
Intelligent Internet Systems - Levy, Weld (2000) (16 citations) (Correct)
....of the task attributes or data from other users. These learning systems have been applied to tasks such as meeting scheduling [31] email processing [87] netnews ltering [109, 16] web search [76] book and music recommendations [86, 111] intrusion dection [74] and web browsing recommendations [85, 99, 98, 64, 125]. For example, when run on data from a faculty member s scheduling behavior, the Calendar Apprentice [31, 89] might learn rules such as Meetings with undergraduates have duration 30 minutes, and take place in my oce while Meetings with the Dean have duration 60 minutes and take place in the ....
M. Pazzani and D. Billsus. Learning and revising user proles: The identication of interesting web sites. Machine Learning, 27:313-331, 1997.
NAUTILUS - Navigate AUtonomously and Target Interesting.. - Gori, Maggini.. (Correct)
....the user pro le by monitoring his her browsing activity. Some other systems use a server based approach, i.e. the user interacts with the Internet through a special server which monitors his her activity and eventually modi es the page to insert controls for its functionalities. Syskill Webert [13, 12] is an intelligent agent based on a naive Bayesian classi er which determines the interestingness of a page. It uses a feedback from the user to adapt the classi er and it can be initialized by specifying an initial set of keywords. The system is topic oriented and the user is required to specify ....
....of the current page and the statistical distribution of the documents retrieved so far. In order to consider also the signi cance of a word with respect to the user interests, we propose to insert a further term into the TF IDF weight in eq. 1) using the idea of information gain also proposed in [12]. The computed weight depends on the particular user u but allows a more accurate selection of the top ranked words taking into account their power in discriminating between documents that are interesting or not interesting for that particular user. The weight is modi es as follows: w u i = f ....
Michael Pazzani and Daniel Billsus. Learning and revising user proles: The identication of interesting web sites. Machine Learning, 27:313-331, 1997.
Long-Term Learning from User Behavior in.. - Müller, Müller.. (2000) (Correct)
....[7] user judgments have been used to propose new items to users based on items being marked together positively by other users. This has been applied to art images of a museum as well [1] The search for user preferences by giving positive and negative examples for web pages has also been studied [15]. Bayesian networks have been used to nd out if an unknown page might t to the users pro le or not. This supervised learning is out of the scope of this paper as we want to use unsupervised learning techniques to avoid additional work for the user. We also want to learn information for new ....
M. Pazzani and D. Billsus. Learning and revising user proles: The identication of interesting web sites. Journal on Machine Learning, 27:313-331, 1997.
Review and Reappraisal of Adaptive Interfaces: Toward.. - Inspired Paradigms Ling (Correct)
No context found.
PAZZANI, M. and BILLSUS, D. 1997, Learning and revising user proles: the identication of interesting web sites. Machine Learning, 27, 313331.
Investigating Semantic Knowledge for Text Learning - Anupriya Ankolekar Forbes (2003) (Correct)
No context found.
M. Pazzani and D. Billsus. Learning and revising user pro les: The identi cation of interesting web sites. Machine Learning, 27(3):313-331, 1997.
Asymmetric Missing-Data Problems: Overcoming the Lack of.. - Aleksander Kocz And (2002) (Correct)
No context found.
Pazzani, M. and Billsus, D.: 1997, Learning and revising user proles: The identication of interesting web sites, 36 Machine Learning 27, 313331.
An Agent-Based Approach for Integrating User Profile Into.. - Pierre, Kacan, Probst (2000) (Correct)
No context found.
M. Pazzani, D. Billsus, Learning and revising user proles: the identi cation of interesting web sites, Machine Learning 27 (1997) 313 331.
Greedy Recommending Is Not Always Optimal - van Someren, Hollink, Hagen (2004) (Correct)
No context found.
M. Pazzani and D. Billsus. Learning and revising user pro les: The identi cation of interesting web sites. Machine Learning, 27:313-331, 1997.
Combining Naive Bayes and n-Gram Language Models for Text.. - Peng, Schuurmans (2003) (1 citation) (Correct)
No context found.
M. Pazzani and D. Billsus. (1997). Learning and Revising User Pro les: The identi cation of interesting web sites. Machine Learning, 27, 313-331.
Online articles have much greater impact More about CiteSeer.IST at NUS Add search form to your site Submit documents Feedback
CiteSeer.IST at NUS - Copyright Penn State and NEC. Hosted by the School of Computing, National University of Singapore.