Information Technology Reference
In-Depth Information
5.2 The Architecture
Figure 6 describes the architecture of the E-shop system with integrated data mining
components. These components include components to access data, clean data, data
mining components, components to visualize the results of the data mining process
and components for the direct usage of the mined knowledge in the e-shop.
In detail these functional components are:
The user interface client components,
The user modeling components
The data mining components,
The knowledge repository,
The visualization component for the web usage mining and user profiles and
The data base with the different media elements and templates.
The data are collected by on-line forms, by server logs and by cookie logs or java
agents in a history list. This information is given as an input to the data mining
component where it can be used for different purposes. The data can be used to learn
the user model and the user preferences as well as the usage of the website. In the data
mining component are realized data mining methods such as decision tree induction
and conceptual clustering for attribute-value based and graph-structured data.
Decision tree induction requires that the data have a class label. Conceptual clustering
can be used to learn groups of similar data. When the groups have been discovered
the data can be labeled by a group name and as such it can be used for decision tree
induction to learn classification knowledge.
Based on the user model the presentation style and the content of the web site
(adaptive multimedia product presentation) are controlled. Besides that the user model
is used to set up specific marketing actions such as e.g. mailing actions or cross-
selling actions. The results of the webusage mining are used to improve the website
organization as well as for monitoring the impact rate of the advertisement of
particular events. Product models and preferences are used to control the content of
the website. The preferences can be learned based on the user´s navigation data.
Besides that an intelligent dialogue component allows to control the dialogue with the
user
.
The following processes can be handled with these components:
9
Web-Site Administration,
Advertisement,
Marketing and Selling,
Adaptive Multimedia Product Presentation,
Event Recognition, and
Learning Ontology Knowledge.
