20.txt

This complete matlab for neural network

发信人: GzLi (笑梨), 信区: DataMining
标  题: web mining
发信站: 南京大学小百合站 (Thu Jul  4 12:44:29 2002), 站内信件

篇名： Web Mining 
刊名： Data Mining and Knowledge Discovery 
ISSN： 1384-5810 
卷期： 6 卷 1 期 出版日期： 200201  
页码： 从 5 页到 8 页共 4 页 
作者： Kohavi Ron   Blue Martini Software, 2600 Campus Dr., San Mateo, CA
 94403, USA. [e-mail]
 
Masand Brij   Verilytics, 1 Wayside Road, Burlington, MA 01803, USA. bmasand
@verilytics.com
 
Spiliopoulou Myra   Leipzig Graduate School of Management (HHL), Department
 of E-Business, Jahnallee 59, D-04109 Leipzig, Germany. [e-mail]
.de
 
Srivastava Jaideep   Computer Science & Engineering, 4-192 EECS Building,
 200 Union Street SE, University of Minnesota Minneapolis, MN 55455, USA.
 [e-mail]


The ease and speed with which business transactions can be carried out over
 the Web
has been a key driving force in the rapid growth of electronic commerce. 
In addition,
customer interactions, including personalized content, e-mail campaigns, 
online customer
service, and online surveys provide newchannels of communication that were
 not previously
available or were very inefficient. The Web is revolutionizing the way busine
sses
 interact
with each other (B2B) and with each customer (B2C). It has introduced entirely
 newways of
doing commerce, including auctions and reverse auctions, micro-segmented 
offers, dynamic
pricing, and up-to-date content. It also made it imperative for organizations
 and companies
to optimize their electronic business.
Knowledge about the customer is fundamental for the establishment of viable
 e-commerce
solutions. As described by Jeff Bezos, CEO of Amazon.com, and mentioned by
 Joseph Pine
in his book The Experience Economy (Pine et al.), customer experience is 
the key to building
customer loyalty in an on-line store because leaving the store is only one
 click away. Web
mining for e-commerce is the application of mining techniques to acquire 
this knowledge
for improving e-commerce. The use of mining techniques in e-commerce can 
help improve
cross-sells, up-sells, assortments shown, ads shown. In addition, clickstream
 collection
allows for unprecedented measurement of site activities, conversion rates
, and the effect of
action (Kohavi, 2001).
WEBKDD 2000 was the second workshop,1 held in conjunction with the Sixth 
ACM
SIGKDD International Conference on Knowledge Discovery in Databases (KDD)
, dedicated
to the challenges of web mining. In response to call for papers, WEBKDD 2000
received 31 contributions. Each was reviewed by at least three program committee
 members.
Seven submissions were selected for presentation as long papers, and six 
as short
papers reporting on good ideas at a rather preliminary phase. The authors
 of all papers
accepted at WEBKDD 2000 were invited to prepare and submit an extended version
 of
their work for consideration in this special issue. All papers were reviewed
 by at least
three reviewers. Based on their ratings and comments, four papers were selected
. The main
selection criteria was the maturity of the work and its impact.
Being able to analyze click-stream data provides an unprecedented opportunity
 to understand
in detail the process leading up to a buy/not buy decision vs. just recording
 the
final outcome—as is the case with point-of-sale data. Clickstream data is
 over 95% of
all data collected in most large-scale e-commerce environments, and contains
 a wealth of
knowledge embedded in it. Tan and Kumar’s Modeling ofWeb Robot Navigational
 Patterns
addresses the challenging and commercially important problem of separating
 the site visits
of web robots from humans. This is crucial for at least two reasons: (1) 
as competitive
pressures increase, commerce sites would like to block robots that collect
 sensitive information,
and (2) accurate modeling of human users’ e-commerce behavior requires that
 web
robot accesses be filtered out. Berendt’s paper, titled Web Usage Mining
, Site Semantics,
and the Support of Navigation, provides a general overview of the issues 
in click-stream
analysis, and shows how the mined knowledge can be used for supporting site
 navigation.
Mobasher, Dai, Luo, Nakagawa, Sun, and Wiltshire’s paper, titled Discovery
 of Aggregate
Usage Profiles for Web Personalization, describe how usage data from web 
logs can be
analyzed/mined to build user profiles, and how these could be use to enhance
 the user’s
browsing experience. Lin, Alvarez and Ruiz’s Collaborative Recommendation
 via Adaptive
Association Rule Mining describe an approach to using association rules for
 collaborative
recommendation. An algorithm for mining association rules, which uses charact
eristics
 of
a target user to improve the efficiency of the mining process, is introduced
.
WEBKDD 2000 turned out to be a very successful workshop. More than 110 people

showed interest in the workshop and over 85 attended it. The quality of papers
 was excellent,
the discussion was lively, and a number of interesting directions of research
 were
identified. This is a strong endorsement of the level of interest in this
 rapidly emerging field
of inquiry. We are pleased to present to the readers the very best papers
 from WEBKDD
2000. Collectively, this represents work that we believe will have significant
 impact on the
future of web mining.2
We wish to thank our reviewers, especially Paul Alpar, Alex Buechner, Jonathan
 Becher,
Michael Berry, Alan Broder, Robert Cooley, Brij Masand, Brad Miller, G¨unter
 Mueller,
Maurice Mulvenna, Carsten Pohle, Myra Spiliopoulou, Jaideep Srivastava, and
 Alex
Tuzhilin.


--
              ***  端庄厚重 谦卑含容 事有归着 心存济物  ***
今天你挖了吗？ DataMining  http://DataMining.bbs.lilybbs.net 
演草纸式的语言 Matlab http://bbs.sjtu.edu.cn/cgi-bin/bbsdoc?board=Matlab

※ 来源:．南京大学小百合站 bbs.nju.edu.cn．[FROM: 211.80.38.29]