CRISP-DM Flashcards
What does CRISP-DM stand for?
Cross Industry Standard Process for Data Mining
What is CRISP-DM?
A document created by Data Mining people which promotes a framework explaining the lifecycle of a DM project.
Created because DM practitioners thought their customers were using DM tools poorly.
1st CRISP-DM step
Business understanding
Description of Business problem/hypothesis and formulation of lower level DM goals to achieve objectives.
2nd CRISP-DM step
Data understanding
Collection of required data and EDA to ascertain key features of the data.
3rd CRISP-DM step
Data preparation
Involves organising and adapting data for the modelling phase
4th CRISP-DM step
Modelling
Different modelling approaches are applied to the data in an attempt to answer DM goals
5th CRISP-DM step
Evaluation
Evaluation of any modelling results achieved
6th CRISP-DM step
Deployment
Deployment of successful models - this could also involve presentation of findings to stakeholders - visualisations