Skip to main content

Table 1 Cross-validated density prediction accuracy of PNN and CPNN models. Average KL divergence for probabilistic and cascaded probabilistic neural network models predicting continuum 3- and 8-class secondary structure from PSI-BlAST encoded sequence data. Window size and number of hidden nodes are varied. All predictions are for 10-fold cross-validation on the training set (set-174). When standard errors are given in parentheses, the predicted value is the mean of five randomized repeats of cross-validation. in parenthesis.

From: Prediction of protein continuum secondary structure with probabilistic models based on NMR solved structures

number of classes window size PNN hidden nodes CPNN hidden nodes
   0 5 10 15 20 25 30 30
3 11 0.59 0.57 0.53 0.52 0.52 0.52 0.51  
  13 0.58 0.57 0.53 0.52 0.51 0.51 0.51  
  15 0.58 0.56 0.53 0.52 0.51 0.50 0.49
(0.002)
0.47 (0.002)
  17 0.58 0.56 0.54 0.52 0.51 0.51 0.51  
  19 0.59 0.56 0.54 0.52 0.51 0.52 0.51  
8 11 0.97 0.97 0.94 0.91 0.92 0.89 0.90  
  13 0.97 0.96 0.92 0.90 0.90 0.89 0.89  
  15 0.97 0.95 0.92 0.90 0.90 0.89 0.88
(0.001)
0.84 (0.002)
  17 0.98 0.96 0.93 0.91 0.90 0.89 0.89  
  19 0.98 0.97 0.94 0.91 0.92 0.89 0.90