peopleanalytics-regression-book
peopleanalytics-regression-book copied to clipboard
keithmcnulty
Chapter 9 questions & solutions
I've attached here my solutions to Chapter 9 and a few questions below:
-
In the Cox proportional hazard model, should we always expect that a single extra point and a single less point of input variables would increase/decrease the odds of outcome with a different percentage?
-
In real-life settings, the headcounts are usually changing over time with new employees joining just as current ones leaving - are there any practical ways for us to conduct similar analysis but with a dynamic sample size?
-
I found that the residuals on intention are not independent of time in the model and therefore the assumption fails to hold. Is this the right result? I did not see a very clear pattern over time in the residual plot though.
