I was talking with an university here in Brazil that is interested in using Moodle prediction systems to help them with avoiding students dropping out of their undergraduate distance courses.
In this case, and in many other cases accross the country, the concept of "course" in Moodle reflects a "class" or an "class offer" or yet "discipline offer", in a term, for example, "Pre-calculus, 2018, first term" or "Linear Algebrar, 2018, second term". With this approach, the undergradutate course, let's say, Computer Science, is represented as a top-level category in Moodle.
The course curriculum could, of course, be represented as a "Learning Plan" , linking all "courses" or "classes" together in a meaningful path to become a computer scientist (in our example).
With the current Moodle's Analytics Models, in this scenario, we can only predict if the student will drop out of a single class, but it would be better if would prediction if he/she will drop out of the course itself, not single classes.
With this set, I ask you: is there any effort already being done to apply the" Students with risk of dropping out" model to Learning Plans?