Skip to main content

Table 4 Classification results for motif-oriented data using different kernels

From: Automatic detection of exonic splicing enhancers (ESEs) using SVMs

  accuracy specificity sensitivity correlation
SVM, combined oligo kernel 90.74% 96.04% 82.09% 78.93%
25% quantile 90.45% 95.4% 81.16% 78.42%
median 90.82% 96.04% 82.09% 79.23%
75% quantile 91.22% 96.62% 83.25% 79.93%
SVM, locality improved kernel 70.00% 92.45% 33.36% 32.43%
25% quantile 69.16% 89.06% 24.45% 30.73%
median 69.88% 91.43% 38.56% 32.33%
75% quantile 70.93% 96.49% 41.37% 34.49%
Markov chain model 68.42% 79.26% 50.71% 31.44%
25% quantile 67.98% 76.17% 50.95% 30.66%
median 68.29% 77.61% 53.7% 31.67%
75% quantile 68.89% 80.57% 55.02% 32.64%
  1. The mean values, 25 percent quantile, median and 75 percent quantile of the accuracy, specificity, sensitivity and Matthews correlation over 50 trials are given.