Information Technology Reference
In-Depth Information
Marketing manager have found that customers who buy oil in a supermarkt will also
by vegetables. Such information can help to arrange a supermarket so that customers
feel more attracted to shop there.
To discover which HTML documents are retrieved in connection with other
HTML documents can give insight in the user profile of the website visitors.
4.3.5 Segmentation
Suppose we have mined a marketing data base for user profiles. In the next step we
want to set up a mailing action in order to advertise a certain product for which it is
highly likely that it attracts this user group. Therefore, we have to select all addresses
in our data base that meet the desired user profile. By using the learnt rules as query
to the data base we can segment our data base into customers that do not meet the user
profile and into those that meet the user profile. The separation of the data into those
data that meet a given profile from those that do not met a given profile is called
segmentation. It can be used for a mailing action where only the address of the
customers who meet a given profile are selected and mailed out an advertising letter.
5 Intelligent E-marketing with Data Mining and User-Adapted
Interfaces
5.1 Objectives
In the following we describe the recent work we are developing for an on-line sales
and advertisement model methods and processes for integrated Data Mining and the
ensuing user-specific adaptation of the web contents. The result shall be tools
comprising the following essential steps:
Identification and recording of web data that are in the following steps the base
for building up user profiles.
Analysis of the data by data mining processes in order to discover and build up
user profiles, e.g. correlation of activities like purchase of associated products or
combining the purchase of some products with certain delivery options.
Integration and visualization, respectively, of the analyzed data for the Web
Content Management and the author process, respectively.
Conversion of the user profiles and the set of rules into user-adapted multimedia
presentations and interaction forms.
The solutions achieved in this project are to be new Data Mining processes, oriented
on the Web-Mining, and allowing at the same time the feedback according to the
knowledge drawn from the Data Mining process to the web contents and the Content
Management, respectively. In the framework of this project we want to investigate
among other things the connection of temporary user modeling with long-term
models, as well as the reciprocal influence of both models in different application
contexts and we want to realize them with the help of component technologies. As a
result there will be prototypical software components at disposal.
