machine learning - Naive Bayes Classifier Biased Output? -
i'm using emgu cv implement machine learning technique in c# classify pixels of image 3 different categories.
everything works perfect far, problem is automatic. want make semi-automatic mean user can "give weight" each of 3 outcomes. give user ability well-tune outcome.
any idea how?
the first thing can think of modify input in way have bias 1 of outputs (for example make more red modifying red channel). though maybe there generic way of doing i'm not aware of.
thanks.
usually you'd adapting prior probabilities in classification rule (what gaussian distributions likelyhood), seems implementation in emgucv not allow that.
Comments
Post a Comment