Table 3 Time-averaged magnitudes of statistical parameters, used to quantify errors in RF model’s predictions of cumulative heat release and heat flow rate.

From: Machine learning enables prompt prediction of hydration kinetics of multicomponent cementitious systems

ML model

R

R2

MAE

MAPE

RMSE

Unitless

Unitless

mW. gCem−1

%

mW. gCem−1

Heat flow rate

RF

0.952

0.912

0.410

23.09

0.63

 

Unitless

Unitless

J. gCem−1

%

J. gCem−1

Cumulative heat release

RF

0.991

0.982

10.45

19.15

14.90