Skip to main content

Table 3 Comparison of the prediction performance for three machine learning algorithms on old data set.

From: NBA-Palm: prediction of palmitoylation site implemented in Naïve Bayes algorithm

Algorithm Window length 3-fold cross-validation 8-fold cross-validation Jack-Knife validation Average MCC Max MCC difference
   Ac (%) Sn (%) Sp (%) MCC Ac (%) Sn (%) Sp (%) MCC Ac (%) Sn (%) Sp (%) MCC   
Naïve Bayes 1 3 85.25% 54.39% 94.21% 0.5438 86.01% 55.98% 94.72% 0.5679 86.33% 57.42% 94.72% 0.5795 0.5637 0.0357
  4 85.97% 56.46% 94.54% 0.5677 86.29% 56.94% 94.81% 0.5776 86.44% 58.85% 94.44% 0.5851 0.5768 0.0174
  5 85.86% 58.53% 93.80% 0.569 86.11% 58.53% 94.12% 0.5757 86.22% 58.37% 94.31% 0.5783 0.5743 0.0093
  6 85.93% 60.13% 93.43% 0.5744 86.58% 62.68% 93.52% 0.5967 86.98% 64.11% 93.61% 0.6099 0.5937 0.0355
  7 86.15% 60.61% 93.56% 0.5811 86.19% 60.61% 93.61% 0.5820 86.44% 61.72% 93.61% 0.5909 0.5847 0.0098
  8 86.01% 60.29% 93.47% 0.5766 86.08% 60.13% 93.61% 0.5781 86.01% 62.20% 92.92% 0.5811 0.5786 0.0045
RBF Network 2 3 84.39% 49.92% 94.40% 0.5104 85.15% 53.91% 94.21% 0.5399 83.85% 50.72% 93.47% 0.4975 0.5159 0.0424
  4 85.11% 53.59% 94.26% 0.5382 85.58% 55.18% 94.40% 0.5543 86.11% 57.89% 94.31% 0.5745 0.5557 0.0363
  5 85.97% 57.10% 94.35% 0.569 85.61% 57.26% 93.84% 0.5596 85.36% 57.42% 93.47% 0.5534 0.5607 0.0156
  6 86.04% 57.42% 94.35% 0.5716 86.11% 59.01% 93.98% 0.5767 85.79% 59.33% 93.47% 0.5689 0.5724 0.0078
  7 85.36% 58.53% 93.15% 0.556 85.65% 57.89% 93.70% 0.5620 86.65% 58.85% 94.72% 0.591 0.5697 0.0350
  8 85.07% 57.26% 93.15% 0.5456 85.43% 57.10% 93.66% 0.5545 86.76% 58.85% 94.86% 0.594 0.5647 0.0484
SVM 3 3 84.46% 56.62% 92.55% 0.5285 85.15% 56.14% 93.56% 0.5448 86.44% 59.81% 94.17% 0.5869 0.5534 0.0584
  4 83.82% 59.97% 90.74% 0.5229 84.03% 58.37% 91.48% 0.5231 82.24% 52.15% 90.97% 0.4616 0.5025 0.0615
  5 82.99% 59.81% 89.72% 0.5041 83.24% 59.97% 90.00% 0.5100 80.52% 53.11% 88.47% 0.4272 0.4804 0.0828
  6 83.10% 62.52% 89.07% 0.5156 83.75% 63.48% 89.63% 0.5326 82.13% 60.77% 88.33% 0.4893 0.5125 0.0433
  7 81.45% 61.72% 87.18% 0.4793 83.21% 63.48% 88.94% 0.5212 83.53% 63.16% 89.44% 0.5269 0.5091 0.0476
  8 80.95% 61.88% 86.48% 0.4702 82.10% 63.16% 87.59% 0.4974 82.45% 63.16% 88.06% 0.5046 0.4907 0.0344
  1. 1. Default parameters of WEKA software package were used.
  2. 2. Ridge value was set to 10E-8.
  3. 3. Polynomial kernel, exponent was set to 1, complexity value C to 0.1.