Interpretable insights

interpretabel insights from predictive algorithms

GOAL

research and evaluate techniques and visualizations to get interpretable insights from predictive algorithms

CONTEXT

  • predictive algorithms often do not provide insights in factors that underlie the prediction
  • approach for discovery of underlying variables and visualization of them

REACH

Within the case study data was used from:

 2015-2016: 1 program, 429 students
2016-2017: 1 program, 409 students

TOTAL: 1 program, 838 students

TARGET AUDIENCE

student advisers and students

PRINCIPLES

actionable feedback

  • insights in predictive algorithms
  • dimensions that matter for student success

return on investment

  • to be explored

social comparison

  • not used

Further reading

project outputs:

scientific output