Context: We are preparing a paper in which machine learning has been used to assign labels to data points. The machine learning algorithm is an ensemble of machines, allowing it to not only output the most likely label, but also to give an estimate of the probability of each of the possible labels being the correct choice. Obviously, the chosen label for each data point will be the label with the highest probability.

Question: we wish to discuss the 'probability' attached to the chosen label (so the max 'probability' for each data point), but that discussion would benefit from adding a name for this property. Can anyone suggest a term which is both fitting, and does not unduly mislead the reader? Obviously the term must be defined and introduced, but still, no point in deliberately confusing the reader.

So far, the best suggestion is 'confidence', since a label with a high estimated probability would, in layman's terms, tend to mean that the ensemble is pretty confident in its decision. However, the point has been made that 'confidence' is usually reserved for discussions related to confidence intervals, which are of course unrelated to this. Personally, I don't think this is a huge issue, since the word would be properly introduced, and our usage is closely related to the 'normal' meaning of the word, but perhaps there is an even better candidate?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.