720.txt

来自「This complete matlab for neural network」· 文本 代码 · 共 61 行

发信人: soullion (river), 信区: DataMining
标  题: the WebWatcher Project 
发信站: 南京大学小百合站 (Wed Feb 27 15:46:49 2002)

"tour guide" agent for the world wide web
Welcome to the WebWatcher Project 


Overview
WebWatcher is a "tour guide" agent for the world wide web. Once you tell it wh
at kind of information you seek, it accompanies you from page to page as you b
rowse the web, highlighting hyperlinks that it believes will be of interest. I
ts strategy for giving advice is learned from feedback from earlier tours. 

WebWatcher Instances
WebWatcher can help you search for information starting from any of the follow
ing pages, but it has learned the most for the first of these. Currently WebWa
tcher is online only on an irregular basis. You might want to take a look at t
he following demo instead. 
CMU School of Computer Science Front Door (modified copy) After arriving at th
is page, click on "WebWatcher" in the top section of the page. 
Machine Learning Information Services 
ARPA Intelligent Integration of Information Home Page 
ARPA Real Time Planning and Control Home Page 
Personal WebWatcher
WebWatcher gives tours to many people (over 8,500 thus far), and learns to bec
ome a specialist at a particular web site. In contrast, our Personal WebWatche
r project stays with a single user, becoming a specialist in that user's inter
ests. 
Publications
WebWatcher: A Tour Guide for the World Wide Web , T. Joachims, D. Freitag, T. 
Mitchell, Proceedings of IJCAI97, August 1997 (longer version internal CMU tec
hnical report September 1996). 
Abstract: We describe WebWatcher as a tour guide agent for the web, the learni
ng algorithms used by WebWatcher, experimental results based on learning from 
thousands of users, and lessons learned from this case study of tour guide age
nts. 
WebWatcher: A Learning Apprentice for the World Wide Web , in the 1995 AAAI Sp
ring Symposium on Information Gathering from Heterogeneous, Distributed Enviro
nments, Stanford, March 1995.
Abstract: A description of WebWatcher and a comparison of different machine le
arning approaches to suggest hyperlinks. 
WebWatcher: Machine Learning and Hypertext, Fachgruppentreffen Maschinelles Le
rnen, Dortmund, Germany, August 1995.
Abstract: A description of how hypertext structure can be used to cluster WWW 
pages. 
Project Members
Dayne Freitag 
Thorsten Joachims 
Tom Mitchell (faculty principle investigator) 
Dunja Mladenic 
Alumni: 
Robert Armstrong 
Ada Lee 
Shannon Mitchell 
David "Stork" Zabowski 
Please report bugs and comments to [e-mail]. 
--

※ 来源:．南京大学小百合站 http://bbs.nju.edu.cn [FROM: 202.112.59.146]