How to practically use ML?

To learn how to practically use machine learning, let’s take the help of a popular framework that depicts a generic data science process.

Data Science Generic Workflow

In the center is data. And around the data are the steps that a modeler does to achieve their goal.

  1. Business Understanding – starts with understanding the problem
  2. Data Understanding – catalogue & collect data
  3. Data Preparation – prepare data for modeling
  4. Modeling – use algorithms to make models
  5. Model Evaluation – evaluate the performance of models
  6. Model Governance – have the model reviewed for risks
  7. Model Deployment – deploy the model as a prediction service

You will note that Model Governance (step 6) is not depicted in such generic diagrams. It is an extra step in mlOS to ensure that the models are fit-for-purpose.

Take note that the data science process is iterative – the modeler goes back and forth many times to perfect their solution.

We will cover these one-by-one in the next few lessons.