Descriptive, predictive, and prescriptive are applicable to both statistics and machine learning.
Descriptive problem:
- Tries to describe the data w/o trying to build any predictions/models into it.
- Eg: Where am I losing my sales?
- it describes the problem
- it tells where the problem is located
- helps isolate/locate where the problem is located
- and requires skill in itself because you need to know how to look at the data.
- Descriptive analytics is trying to understand certain things about data that help me get to conclusions about the data.
- It is simply telling you what is there WRT the question that you ask of it.
predictive problem:
- if I increase the price of the watches(into a luxury brand), will my sales increase?
- it says to look at this data and give me an idea as to what might happen
- what would happen if I change this/that?
Prescriptive problem:
- How do I increase the sales without losing out on the overall bottom line per sale?
- A prescription typically has to satisfy multiple requirements.