The best thing to do is compare your results to a widely known dataset with information about the accuracy rate for what you implemented. That will give you a b

The best thing to do is compare your results to a widely known dataset with information about the accuracy rate for what you implemented. That will give you a baseline to figure out what, if anything, is going wrong. > Wrong algorithm? Almost certainly not, Bayes is a very solid classification technique.