Looking at the information flow of the decision tree, some pandas like big butts, and some do not. Most other decisions in the tree appear to be inconsequential. Tallying up the score, it turns out, your decision tree isn’t much better than the previous algorithm according to your test set. It’s better, so you could get away with it. Maybe panda sexuality is just one of those areas best left untread by science. Or maybe, this is just the best you’re going to get. Or maybe, you can supervisor of the quarter award winning motivate this algorithm into something fantastic! What do you do?
Add in Big 5 Personality Trait Test, if it works for humans, why not pandas
Add in Dating Questionnaire from some magazine you accidentally stole from the dentist’s office.
Add in more physical traits, hey, that’s how you work right?