Evaluating machine learning for projecting completion rates for VET programs

This technical paper evaluates the effectiveness of using machine learning approaches to calculate projected completion rates for [vocational education and training] VET programs compared with the current Markov chains methodology. Projected rates were calculated for the 2016 commencing cohort with completion rate predictions using machine learning algorithms generally more accurate than the rates achieved using Markov chains methodology. While the machine learning model results show how well the model can generate projected rates for the 2016 commencing year, it is untested how well the...  [+] Show more

This technical paper evaluates the effectiveness of using machine learning approaches to calculate projected completion rates for [vocational education and training] VET programs compared with the current Markov chains methodology. Projected rates were calculated for the 2016 commencing cohort with completion rate predictions using machine learning algorithms generally more accurate than the rates achieved using Markov chains methodology. While the machine learning model results show how well the model can generate projected rates for the 2016 commencing year, it is untested how well the model is able to consistently make accurate predictions for other commencing years and if assumptions underlying either methodology remain valid for years where training may have been disrupted by the [Coronavirus Disease 2019 (COVID-19)] pandemic.

Summary from publisher's website.