bayesian

bayesian copied to clipboard

LogScores, ProbScores, and SafeProbScores all have a return parameter that is the index of the most likely class. I think if you're ever willing to break the current api it would be a ton more useful to return the actual bayesian.Class.

It would make simple usage as below much easier.

As it stands I have to know the index of which I passed it into my classifier. That's knowledge I'd rather not have to know. It's kind of a difficult way to do it.

const (
    Good bayesian.Class = "Good"
    Bad  bayesian.Class = "Bad"
    Ugly bayesian.Class = "Ugly"
)

classifier := bayesian.NewClassifier(Good, Bad, Ugly)

_, c, _, _ := classifier.SafeProbScores(wht)

if c == Ugly {
    fmt.Println("oh no")
}

Fair point, all. Will look into it.

@donatj

func (c *Classifier) GetClass(intx int) string { return string(c.Classes[intx]) }

@mycalf I don't see that in the API.

func (c *Classifier) GetClass(intx int) Class { return c.Classes[intx] }

Returning the actual Class would be more useful than string, but to me it would still make far more sense for the API to return the class rather than the class index throughout.

After reviewing:

I'm recalling that the reason I am returning the scores and an index is because different applications may wish to examine the actual scores across all potential classes. If I return a Class and not an index, then I also have no (ready) way of determining what the actual score was for the most likely class. So, some solutions:

1. Return the index and the class. (Starting to overly crowd in the return prototype, IMO.)
2. Add an auxiliary GetClass function as @mycalf suggests. (Clunky.)
3. Return a map which maps classes to scores. (More memory intensive.)
4. Add helper methods like MostLikelyClass() to encapsulate the complexity of looking up the class if the scores are not being considered. (My recommendation.)

Thoughts?

MostLikelyClass() seems like a good compromise..