0

I am new to ML and currently using Python. If I am testing a set of interventions on a certain outcome . What kind of tools/Python programs should I start learning to begin to determine which intervention (or combination of interventions) is most likely to predict a certain outcome.

Here is an example: I want to test which interventions promote recycling behavior most effectively (outcome is 'yes' or 'no' recycling)...different combinations of interventions will be used (i.e phone call, email reminder, text prompt, etc). I want to determine which of these interventions (or which combination of interventions) was most successful at promoting recycling

Thanks everyone and stay healthy!

Jco15
  • 1
  • 1
  • Can you give a bit more context and a simple example? – Thom Ives Apr 09 '20 at 16:23
  • 1
    thanks for the responses here is an example: I want to test which interventions promote recycling behavior most effectively (outcome is 'yes' or 'no' recycling)...different combinations of interventions will be used (i.e phone call, email reminder, text prompt, etc). I want to determine which of these interventions (or which combination of interventions) was most successful at promoting recycling – Jco15 Apr 09 '20 at 16:27
  • If you restructure your question with you reply in the comments along with what you've learned in the answer provided by David Smolinski, your question will be more helpful to others looking for similar help. – Thom Ives Apr 09 '20 at 16:39

1 Answers1

1

"Dimensionality reduction" is the ML topic of reducing the number of features used to predict a response.

Finding correlation with pandas.DataFrame.corr helps.

David Smolinski
  • 514
  • 3
  • 13
  • 1
    So if I have n number of interventions, this approach would help determine the best intervention (or combination of interventions) to achieve a desired outcome? Thanks and sorry for the naive questions. – Jco15 Apr 09 '20 at 16:22