Skip to main content
. 2015 Dec 1;3(4):230–237. doi: 10.1089/big.2015.0019

FIG. 3.

FIG. 3.

The four heatmaps show (A) the root mean square error (RMSE), (B) the coefficient of determination (R2), (C) the number of points without prediction, and (D) the number of layers in the resulting casade, when the number of predictors in each layer (m) and the data subset size (p) are varied. Our requirements for the combination of m and p were a low test RMSE, high R2, with reasonable number of points without prediction that go to the experts, and fewer layers.