Skip to main content

Table 2 Structural classification performance using KNN for the Full-SCOP dataset

From: Cutoff Scanning Matrix (CSM): structural classification and function prediction by protein inter-residue distance patterns

SCOP Level

Before SVD

After SVD

∆Prec.

∆Rec.

 

Precision

Recall

Precision

Recall

  

Class

0.927

0.926

0.954

0.954

+2.7%

+2.8%

Fold

0.868

0.869

0.922

0.923

+5.4%

+5.4%

Superfamily

0.871

0.872

0.926

0.927

+5.5%

+5.5%

Family

0.888

0.889

0.938

0.938

+5.0%

+4.9%

  1. Prediction performance for the full-SCOP dataset using KNN. The experiment was performed for each classification level of SCOP. The precision and recall metrics are weighted averages. A 10-fold cross validation was employed.