New answers tagged

1

I'm not sure of any literature, as its a pretty specific question, but I tried my hand at a very quick and dirty simulation study. For 1000 cases, I have 10 variables. Each of these 1000 cases has two instances of the same variable, such as we'd see over time. The correlation is either r = .0, .25, .5, .75, or .99. I pivot the data into tidy format, such ...


0

If you want to find a linear boundary, linear discriminant analysis (LDA) is a good choice Wikipedia page on 'Linear Discriminant Analysis'. Logistic regression would also probably yield a similar result. You can add polynomial features to either of these methods to find nonlinear boundaries, but interpreting the results could get more complicated. I think ...


Top 50 recent answers are included