Making predictions with Moodle Analytics
Custom and log-based reports are descriptive; they tell viewers what happened but not why, and don’t predict outcomes or advise users on how to improve. While reporting tells us about who, what, when, and where, learning analytics aims to explain why and how well.
Moodle Analytics predicts or detects unknown aspects of learning based on (historical) log data and current behavior.
There are two types of models that Moodle Analytics supports:
- Machine learning-based models: Sophisticated models use mathematical tools to make predictions; for instance, students with a high likelihood of failing a course
- Static models: Simple models are based on assumptions and do not require sophisticated analyses; for example, students who have not enrolled in a course
The following diagram illustrates the...