Table 2 Performance metrics for the regression analysis

From: Unravelling the human taste receptor interactome: machine learning and molecular modelling insights into protein-protein interactions

 

RMSE

RMSE (ΔG)

RAE

RRSE

Spearman correlation

Spearman p value

Training total

0.0951 ± 0.005

8.12 ± 0.42

86.03% ± 5.38%

83.82% ± 4.67%

0.81 ± 0.108

1.56e−60

CV total

0.1082 ± 0.0008

9.24 ± 0.07

97.78 ± 0.88%

98.72% ± 0.67%

0.22 ± 0.02

3.28e−05

Ensemble testing

0.1233 ± 0.0004

10.53 ± 0.03

99.69% ±0.72%

98.54% ± 0.41%

0.19 ± 0.01

0.05 ± 0.01

  1. Metrics are provided for cross-validation (CV), training, and testing sets. RMSE values are provided for both the normalized ΔG values of the endpoint (0–1 range) and for absolute ΔG values (RMSE (ΔG)). Ensemble testing references to the ensemble testing method based on majority voting of the Pareto Front models.
  2. RMSE root mean square error, RAE relative absolute error, RRSE root relative squared error.
Back to article page