2019 SIMON INITIATIVE LEARN LAB, CARNEGIE MELLON UNIVERSITY
DATA CLEANSING - We cleaned and managed student problem-solving log data of a geometry tutor from Datashop @ CMU
KC MORDEL OPTIMIZATION- Based on the original model by Kenneth Koedinger & John Stamper, we built up a new Knowledge Component Model by splitting trapezoid-related problem steps from the find_individual_area knowledege component steps.
MODEL EFFICACY - Our Knowledge Component Model decreased RMSE by 0.2%, and indicated that it is nessesary to split trapizoid-repated steps from find_individual_area knowledge component.
