I am trying to build a model using logistic regression, where my dependent variable is y=1 if the mail was opened, y=0 if it was not.

I have data approximately 10 records (10 rows) for every recipient who got an email and want to calculate the probability of opening the email per user. However, do not know how to edit the dataset based on this requirement.

Because the results (probabilities) now are like for every row (every mail) and I want it for every recipient.

ID is an ID of the recipient(see the screenshot) If I counted the number of openings, I would lose binary operator, what is necessary for logistic regression.

Got any ideas?

enter image description here

  • $\begingroup$ Can you explain why you want to apply logistic regression? A naive approach to this would be to just count the empirical probabilities per user, e.g. $\hat{p}(OPENED=1 \mid ID = 1) = 7/14 = 0.5$. $\endgroup$ – Sammy Jan 13 at 8:09
  • $\begingroup$ You could predict the % chance of a particular user opening an email. I.e. user 1 opens 3 of 10 emails their response variable is y = 0.3. If you have other columns in this dataset (more info on the emails) this would be a good place to apply decision trees. Hard to say exactly without more info - are there more columns in your dataset? $\endgroup$ – bstrain Jan 15 at 1:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.