Skip to main content

Advertisement

Table 4 Size of the training dataset vs. execution times taken by the Prote2S and the SVM during the training process.

From: Prediction of protein secondary structures with a novel kernel density estimation based classifier

  Prote2S SVM
Number of protein chains used to generate the training dataset Training time (in seconds) Q3 SOV Training time (in seconds) Q3 SOV
50 29.6 64.0% 52.9% 138.08 71.3% 64.3%
100 91.7 69.0% 64.1% 527.02 74.0% 68.3%
250 486.4 71.4% 67.2% 5105.63 75.5% 71.0%
500 1377.4 71.9% 67.9% 21040.0 76.8% 72.3%
1000 3887.8 73.9% 71.1% 78795.25 77.4% 73.3%