Table 3

Performance comparison of different approaches using cross-validation for the benchmark data sets.

PS1302

Localization
 HYBRID
 CELLO
 PSORTb v.1.1
 PSLpred
 P-CLASSIFIER


Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
CP
 95.6
 0.90
 90.7
 0.85
 69.4
 0.79
 90.7
 0.86
 94.6
 0.85
IM
 93.3
 0.94
 88.4
 0.92
 78.7
 0.85
 86.8
 0.88
 87.1
 0.92
PP
 91.4
 0.88
 86.9
 0.80
 57.6
 0.69
 90.3
 0.90
 85.9
 0.81
OM
 96.3
 0.96
 94.6
 0.90
 90.3
 0.93
 95.2
 0.95
 93.6
 0.90
EC
 90.0
 0.89
 78.9
 0.82
 70.0
 0.79
 90.6
 0.84
 86.0
 0.89
Overall
 93.7
 -
 88.9
 -
 74.8
 -
 91.2
 -
 89.8
 -

PS1444

Localization
 HYBRID
 CELLO II
 PSORTb v.2.0
 PSLpred
 P-CLASSIFIER


Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
 Acc (%)
 MCC
CP
 95.0
 0.91
 95.3
 0.89
 70.1
 0.77
 -
 -
 -
 -
IM
 93.5
 0.94
 90.0
 0.91
 92.6
 0.92
 -
 -
 -
 -
PP
 90.2
 0.86
 87.7
 0.82
 69.2
 0.78
 -
 -
 -
 -
OM
 96.7
 0.96
 92.8
 0.90
 94.9
 0.95
 -
 -
 -
 -
EC
 87.4
 0.87
 79.5
 0.82
 78.9
 0.86
 -
 -
 -
 -
Overall
 93.2
 -
 90.0
 -
 82.6
 -
 -
 -
 -
 -

§ The best performance of overall and individual localization sites is underlined.

Su et al. BMC Bioinformatics 2007 8:330   doi:10.1186/1471-2105-8-330