I am working on electricity theft detection. I am considering Electricity consumption history for the past 24 months as a feature. Along with this information, I also have consumers' payment history, whether they are paying their electricity bill on time. I have taken the payment history feature as a count of months in which they haven't paid the total bill. How can I be confident that I should select the count of months as a feature for my ML model?

  • $\begingroup$ Why is this question tagged as "hypothesis-testing"? Concerning feature selection, you might have a look at boruta and random forests, see, e.g., this thread: stats.stackexchange.com/q/264360/244807 $\endgroup$
    – cdalitz
    Jan 5 at 11:44


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.