Skip to main content

Table 4 The average correlation coefficient (CC) and RMSE for the PD6 and PD7 datasets.

From: Improved estimation of structure predictor quality

 

PD6

PD7

Method

CC

RMSE

CC

RMSE

Custom training

   Support Vector Regression

0.87

3.33

0.89

2.86

   Linear Perceptron

0.87

3.81

0.89

2.93

   Standard Regression

0.70

6.02

0.82

3.80

   Constrained Regression

0.87

3.46

0.89

2.86

Global training

   Support Vector Regression

0.90

2.97

0.90

2.67

   Linear Perceptron

0.88

3.64

0.89

2.90

   Standard Regression

0.89

3.20

0.90

2.76

   Constrained Regression

0.89

3.15

0.89

2.79

Static consensus

   LGA-Distance

0.86

4.29

0.89

3.19

   LGA-S-score

0.86

3.74

0.83

3.56

   LGscore-Distance

0.67

5.56

0.72

4.48

   LGscore-S-score

0.81

4.57

0.79

4.33

  1. Values in this table represent the Pearson correlation coefficient (CC) and root mean squared error (RMSE) between the predicted and true per-residue distances. Boldfaced entries correspond to the best performing scheme for each dataset and performance assessment metric.
Back to article page

BMC Structural Biology

ISSN: 1472-6807

Contact us