Fig. 6From: Enhancing data pipelines for forecasting student performance: integrating feature selection with cross-validationAverage percent difference between the AUC values obtained using and omitting feature selection. In each data table entry, the first number denotes the mean AUC, across all training and testing corpora, when preprocessing feature selection are applied, and six features are used as input to the DMM. The entries in parenthesis correspond to the mean percent difference in the AUC, denoted by the colored palettesBack to article page