Influence of Data Splitting on Performance of Machine Learning Models in Prediction of Shear Strength of Soil

<div>Probability density results for 1000 Monte Carlo simulations for the testing part with different indicators: (a) RMSE; (b) MAE; and (c) <i>R</i>. It should be noticed that the legends, for instance, Train 10%, indicates that the results obtained with 10% of the total data were used to construct the training dataset.</div>

Mathematical Problems in Engineering

fig4

Figure 4

Figure 4: Influence of Data Splitting on Performance of Machine Learning Models in Prediction of Shear Strength of Soil 