CRISP-DM process: the six phases.
As illustrated in Figure 3-1, CRISP-DM presents the process of a
data mining project through a life cycle with six distinct phases. It high-
lights the tasks associated with each phase as well as the relationships
between the phases and tasks. As with most standards, CRISP-DM
does not claim to cover every relationship between phases and tasks,
since this depends on project goals, the user's experience and needs,
and the peculiarities of the data. It is highly likely that movement
between any of the defined phases may be required. The arrows in the
figure indicate common or the most important phase relationships.
The data mining process itself typically forms a continuum as
indicated by the outer circle of Figure 3-1. Once a solution has been
deployed, new insights into the problem typically emerge, yielding
more questions that can be answered by data mining or refinements
of the existing solution to improve result quality. With each iteration
of the data mining process, improved skills and experience help
improve subsequent efforts.
Business Understanding Phase
In the first phase, CRISP-DM begins with the problem to be solved,
called business understanding, which includes defining business