classifiers

Small programs generated using Machine Learning techniques. The general idea is that instances of past student behavior, either individual or in groups, are annotated by teachers, moderators, or pedagogical experts as belonging to certain categories of interest, such as displaying critical reasoning (or not), critical analysis of opinions (or not), on-topic-conversation (or not), etc., and then the Machine Learning algorithms learn from these labeled instances, generating classifiers for each of the categories of interest.

(from: M4.1, by Bruce M. McLaren? & Oliver Scheuer, DFKI)