CRISP-DM
Cross industry standard procedure for data mining: A friend recommended this to me as a decent methodology for considering data science / mining projects. Although it's fairly old now (pre-2000), it has some useful concepts which you can use to articulate the stages of such a project (and to make sure you're not missing any...).
Broadly it uses the following diagram to show the overall process:
Click permalink to read more...
It then goes on to define each of these phases and break them down into a number of 'generic tasks', like the following:
The methodology also goes on to explain each of these tasks, with proposed activities and outputs for each, as well as best-practices and common pitfalls.
There are several different iterations, but these diagrams were pulled from:
This one also looks good:
ftp://public.dhe.ibm.com/software/analytics/spss/documentation/modeler/14.2/en/CRISP_DM.pdf
There are no published comments.
New comment