The phases of an engagement model
A very well-known modeling process is CRISP-DM for data mining and predictive analytics projects, and it involves six steps:
- Business understanding
- Data understanding
- Data preparation
- Modeling
- Evaluation
- Deployment
Each of these phases follows each other, and some of them happen in a recursive fashion by providing feedback to the preceding phase. The deployment phase is particularly important in terms of model monitoring and maintenance, which is the focus of this chapter.
Let's quickly look at each of these phases and their purpose in the overall process.
Business understanding
This is the phase where you fully focus on the business objectives in terms of project goals, scope, resources, limitations, iterations, and checkpoints.
You try to understand the whole picture in business terms and frame the problem technically. There might be competing objectives between different internal stakeholders in your organization, and you should be aware of them and find the sweet...
