Research Article

Protein Sequence Classification with Improved Extreme Learning Machine Algorithms

Table 3

Successful testing classification rate and standard derivation (rate (%) ± Dev (%)) comparisons of different classifiers: randomly generated training and testing datasets from the mixed protein sequences, where L, S, and G stand for the linear kernel, the sigmoid kernel, and the Gaussian kernel, respectively.

Methods LSG

VOP-ELM 97.30 ± 0.73 98.19 ± 0.70 98.68 ± 0.71
V-ELM 96.91 ± 0.71 97.75 ± 0.64 97.74 ± 0.59
OP-ELM 95.95 ± 0.29 96.42 ± 0.45 97.55 ± 0.55
SVM 97.17 ± 0.49 97.28 ± 0.63 97.33 ± 0.66
BP 94.93 ± 0.95 96.29 ± 0.85 95.54 ± 0.91
ELM 94.94 ± 0.75 96.72 ± 0.85 96.65 ± 0.63