Developed at Rice University, SPARFA (short for SPArse Factor Analysis), is an analytical tool for large educations datasets where students interact with homework and test questions. SPARFA uses powerful statistical models to decompose educational data as consisting of a small number of educational concepts. From the concept-centric view of the data, SPARFA perform two important educational tasks:
1) Learning Analytics: SPARFA analyzes educational data to determine how well students have mastered various educational concepts.
2) Content Analytics: SPARFA additionally educational test items to determine which concepts are related to various test items.
By performing these tasks, SPARFA automatically enables a number of important educational tasks, such as providing feedback to students, instructors, and educational content authors in order to improve the educational experience.
Over the years, we have organized a couple of workshops at various machine learning/data mining conferences.
KDD 2017 Workshop on Machine Learning for Education: http://ml4ed.cc/
For the upcoming NIPS 2016 workshop on Machine Learning for Education, visit https://dsp.rice.edu/ml4ed_nips2016
Past workshops:
ICML 2016 workshop on Machine Learning for Digital Education and Assessment Systems: http://medianetlab.ee.ucla.edu/ICML-Education2016.html
ICML 2016 workshop on Machine Learning for Digital Education and Assessment Systems: http://medianetlab.ee.ucla.edu/ICML-Education2016.html
ICML 2015 workshop on Machine Learning for Education: https://dsp.rice.edu/ML4Ed_ICML2015
NIPS 2014 workshop on Human Propelled Machine Learning: http://dsp.rice.edu/HumanPropelledML_NIPS2014
Submissions and talk slides can be found on these pages.