🏠

2

3

4

5

6

7

8

Amit Kishore on How to Define a Problem

<aside> <img src="https://s3-us-west-2.amazonaws.com/secure.notion-static.com/4c744779-1d02-4ff5-94b8-03a6441be71a/icons8-google-48.png" alt="https://s3-us-west-2.amazonaws.com/secure.notion-static.com/4c744779-1d02-4ff5-94b8-03a6441be71a/icons8-google-48.png" width="40px" /> Google’s Rules of Machine Learning

</aside>

This is to understand and articulate the business logic of the problem.

Need a problem to solve? Try the recommendations in this video, or this list by industry on GitHub, find something on projectpro.io, or join an Omdena project.

Set expectations:

• How to frame a problem?
• The benefit or ultimate goal that is trying to be achieved
  • Iterative, you may have to come back and refine the goal
  • Characteristics of a Good Question
  • The Six Types of Questions
• The inherent assumptions of the problem and data
• Limitations of data, ability, and technology
• Define success; be specific and measurable.
• Built the team:
  • An iterative process. The initial team may grow or shrink as the requirements are identified. Having the right team members on board as soon as possible will allow a better assessment of the problem and project needs.
  • Are the needed people resources available? If not, what is the timeline before they are?

Explore the question:

• George Polya’s Problem Solving Techniques
• [Statistical Power Analysis](https://www.statisticssolutions.com/dissertation-resources/sample-size-calculation-and-sample-size-justification/statistical-power-analysis/#:~:text=Power analysis is normally conducted,the desired level of significance.) -
  • Statistical Power - Machine Learning Mastery