You Evaluate your Naive Bayes

You start testing out the Naïve Bayes algorithm and you’ve found that the zero frequency problem is incredibly real with the “forever alone” pandas. There would be an enormous subset of pandas who will never have sex. Alternatively, the ones who will are expected to get real busy and produce a modest improvement in panda fertility. You could go on with the algorithm, or you could change up the priors to try and help out the “forever alone” pandas. What do you do.

Good enough, not much you can do about that

Change up the priors on your naïve Bayes to favor pandas with intelligence and less for the physical traits

Change your priors to favor emotional availability as many of the studly bull-pandas taking up all the mates from the “forever alone” pandas will never have a meaningful relationship

Heavily weight your priors on the ability to play DnD, as nearly every classified “forever alone” panda had a propensity for roll playing games and not the sexy type of roll playing games where one panda’s a plumber and the other doesn’t know how to pay for their goods and services and doesn’t accept IOUs

Try a Decision Tree

Use some ole Fashioned Regression

Give into peer pressure and use a Neural Net!