Table 1 The performance of amino acid composition-based models developed using various machine learning techniques for predicting hemotoxicity of peptides.
From: A Web Server and Mobile App for Computing Hemolytic Potency of Peptides
Methods | Sn (%) | Sp (%) | Acc (%) | MCC |
|---|---|---|---|---|
HemoPI-1 main dataset | ||||
SVM | 95.7 | 94.8 | 95.3 | 0.91 |
IBK | 95.5 | 93.7 | 94.6 | 0.89 |
Multilayer Perceptron | 93.9 | 92.8 | 93.3 | 0.87 |
Logistic | 93.4 | 93.7 | 93.6 | 0.87 |
J48 | 89.6 | 88.5 | 89.0 | 0.78 |
Random Forest | 94.1 | 94.6 | 94.3 | 0.89 |
HemoPI-2 main dataset | ||||
SVM | 76.0 | 76.8 | 76.4 | 0.53 |
IBK | 75.6 | 76.0 | 75.7 | 0.51 |
Multilayer Perceptron | 74.0 | 74.1 | 74.0 | 0.48 |
Logistic | 64.5 | 68.1 | 66.1 | 0.32 |
J48 | 79.0 | 60.0 | 70.3 | 0.40 |
Random Forest | 77.8 | 77.8 | 77.8 | 0.56 |
