Answer to "What are predictive learning analytics?"

Topic Editor: Dr. Brent Thoma, University of Saskatchewa

















Predictive learning analytics use historical data to develop models which make inferences about uncertain future events.[1,2] For example, data from a population of learners with known learning outcomes within a training program (e.g. whether they will require remediation, be placed on probation, or fail to complete a training program) can be used to predict these outcomes for a group of learners whose outcomes in the training program are not known. These predictions are based upon known characteristics of the individual members of each cohort that correlate with each outcome of interest.


Analytic techniques used within predictive learning analytics require clear subject characteristics and a large data set of comparable subjects with known outcomes of interest. Techniques include linear and logistic regression, 'nearest neighbors' classifiers, decision trees, Bayes classifiers, Bayesian networks, support vector machines, and neural networks.[1,2] As data on medical learners and their work becomes increasingly plentiful both within learning management systems and electronic medical records it will be increasingly possible to access information that may have predictive value.[3]



1 – Chan T, Sebok‐Syer S, Thoma B, Wise A, Sherbino J, Pusic M. Learning Analytics in Medical Education Assessment: The Past, the Present, and the Future. AEM Education and Training. 2018 Apr;2(2):178-87.

2 – Lang C, Siemens G, Wise A, Gasevic D, editors. Handbook of learning analytics. SOLAR, Society for Learning Analytics and Research; 2017.

3 - Arora VM. Harnessing the Power of Big Data to Improve Graduate Medical Education: Big Idea or Bust?. Academic Medicine. 2018 Jun 1;93(6):833-4.