Before we start modeling
Before collecting data as the starting point of a machine learning life cycle, you need to know your objectives. You need to know what problems you want to solve and then define smaller subproblems that would be machine learning solvable. For example, in the case of a problem such as, “How could we reduce the number of fragile products returned to a manufacturing facility?,” the subproblems could be as follows:
- How could we detect the cracks before packaging?
- How could we design better packaging to protect the products and reduce transportation-caused cracks?
- Could we use better materials to reduce the risk of cracking?
- Could we apply small design changes to our product that do not change its functionality but reduce the risk of cracking?
Once you have identified your subproblems, you can find out how you can use machine learning for each and go through a machine learning life cycle for the defined subproblems. Each...
