Clickstream data is one of the most important sources of information in websites usage and customers' behaviour in e-commerce applications. A number of web usage mining scenarios are possible depending on the available information. While simple traffic analysis based on clickstream data may easily be performed, such techniques are not adequate to substantially improve e-commerce activities. Meaningful results require a more complex research design. Questions of relevant data sources, possible applications and a prototype for web usage mining are briefly outlined in this paper. Aiming for more than simple traffic analysis, we need to develop models and tools to integrate additional databases such as customer data, content structure data and data from OLTP-systems. The design and implementation of such integrated data sources, which may be called data webhouse, is the necessary starting point of any analytical mining technique. The application is demonstrated within various examples.
[1]
Ron Kohavi,et al.
Mining e-commerce data: the good, the bad, and the ugly
,
2001,
KDD '01.
[2]
Jaideep Srivastava,et al.
Web usage mining: discovery and applications of usage patterns from Web data
,
2000,
SKDD.
[3]
Marcus Hudec,et al.
Model-based classification of large data sets
,
2002
.
[4]
Ralf Walther,et al.
The Data Webhouse Toolkit
,
2001,
Künstliche Intell..
[5]
R. Amit,et al.
Value creation in E‐business
,
2001
.
[6]
Lars Kai Hansen,et al.
Webmining: learning from the world wide web
,
2002
.
[7]
Elmer B. Mode,et al.
Elements of Statistics
,
1942,
Agronomy Journal.
[8]
John McCarthy,et al.
Phenomenal data mining: from data to phenomena
,
2000,
SKDD.
[9]
Renato Coppi,et al.
A theoretical framework for data mining: the informational paradigm
,
2002
.