Table 1.
The training dataset used in this work [2].
Table 2.
The whole test dataset [2].
Fig 1.
Methodology framework for correlation of crude oil viscosity.
Fig 2.
Structure of decision tree model.
Fig 3.
Demonstration of Cook’s distance outlier detection.
Table 3.
The comparisons of three developed models.
Table 4.
Final values of R2-scores for three models.
Fig 4.
Actual Vs. estimated outputs (DT model).
Fig 5.
Actual Vs. estimated outputs (GRNN model).
Fig 6.
Actual Vs. estimated outputs (MLP model).
Fig 7.
Residual plot of tuned MLP model.
Fig 8.
Effect of API and ρ (g/cc) on the output.
%C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, MWC7+ = 338.41, SGC7+ = 0.95, T (°C) = 30.
Fig 9.
Influence of SGC7+ and ρ (g/cc) on the output.
API = 18.6, %C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, MWC7+ = 338.41, T (°C) = 30.
Fig 10.
Effect of temperature and ρ (g/cc) on the output.
API = 18.6, %C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, MWC7+ = 338.41, SGC7+ = 0.95.
Fig 11.
Trends of API on multiple temperature levels (%C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, MWC7+ = 338.41, SGC7+ = 0.95, ρ (g/cc) at 20°C = 0.941).
Fig 12.
Trends of MWC7+on multiple temperature levels (API = 18.6, %C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, SGC7+ = 0.95, ρ (g/cc) at 20°C = 0.941).
Fig 13.
Trends of density on multiple temperature levels (API = 18.6, %C4 = 0.35, %C5 = 1.62, %C6 = 3.21, %C7+ = 94.81, SGC7+ = 0.95.