A Reference Model for Learning Analytics

Mohamed Amine Chatti, Onlea, eCOTOOL, DigitalKoans, Jan 22, 2013
Commentary by Stephen Downes
LA_RefModel_Chatti.png

This is a good overview of learning analytics, beginning with the concept of the recommender system as it developed in the 2000s through to more contemporary work. Chatti breaks the discussion of learning analytics into four major areas of investigation:

  • - data and environments, ie., what system is collecting the data - today very little data is collected in personal learning environments and even learning management systems, while a great deal of data is captured by adaptive systems and web-based courses
  • - stakeholders, or, who is collecting the data - today intelligent tutors and researchers or system designers are the primary users of data, while institutions, students and teachers use much less data
  • - methods, or, how is the data collected - statistics and visualizations are widely used, while classification and prediction (holdovers from recommender systems) are used most frequently; upcoming are social network analysis and association rule mining
  • - objectives, or, why is the data being mined - today it's mostly for monitoring and analysis and (in other systems, probably) for intelligent tutoring and adaptation, while other uses unclude intervention, assessment and feedback, and personalization.

Related: Adam Cooper, in A Briaf History of Analytics (via Analytics is Not New) suggests "we would be wise to avoid assuming that analytics implies big data and the latest predictive data mining algorithms... (using analytics) involves a range of options, some of which are lateral rather than evolutionary developments."

Views: 0 today, 128 total (since January 1, 2017).[Direct Link]