Skip to main content

Table 7 Comparison with the state-of-the-art methods for multi-class remote fold recognition

From: Efficient use of unlabeled data for protein sequence classification: a comparative study

Method Error Top-5 Error Balanced Error Top-5 Balanced Error F1 Top-5 F1
mismatch (full seq.) 50.49 22.31 76.44 38.61 24.96 65.58
triple (full seq.) 50.81 17.92 71.95 27.80 28.92 73.93
mismatch (region) 44.79 13.36 67.26 25.40 33.17 77.45
triple (region) 36.81 10.91 52.58 20.07 49.69 81.26
profile(5,7.5) 45.11 15.80 71.27 31.55 32.34 75.68
profile(5,7.5)† 46.30 14.50 62.80 23.50 - -
  1. †: directly quoted from [12]
\