Databases Reference
In-Depth Information
knowledge, and proper interpretation of the results of mining, are also essential
to derive useful knowledge from data.
In our opinion the main problem of the framework presented in Fig. 2 is
that all KDD activities are seen from “inside” of DM having nothing to do
with the relevance of these activities to practice (business).
CRISP-DM: CRoss Industry Standard Process for Data Mining
The life cycle of a DM project according to the CRISP-DM model (Fig. 3) con-
sists of six phases (though the sequence of the phases is not strict and moving
back and forth between different phases normally happens) [6]. The arrows
indicate the most important and frequent dependencies between phases. And
theoutercircleinthefiguredenotesthecyclicnatureofDM-aDMprocess
continues after a solution has been deployed. If some lessons are learnt dur-
ing the process, some new and likely more focused business questions can be
recognized and subsequently new DM processes will be launched.
We will not stop at any phase of CRISP-DM here since it has much over-
lapping with Fayyad's view and with the framework that is considered in the
next section and discussed in more details. However, we would like to notice
that the KDD process is put now in a way into some business environment
that is represented by the business
understanding
and
deployment
blocks.
Business
Understanding
Data
Understanding
Data
Preparation
Deployment
Modelling
Data
Evaluation
Fig. 3.
CRoss industry standard process for data mining [6]
