Skip to main content

Table 3 Performances of miPred and miR-KDE for the NH3350 dataset in terms of genus.

From: Using a kernel density estimation based classifier to predict species-specific microRNA precursors

 

%SE

%SP

%ACC

%Fm

%MCC

Vertebrata

   miPred

95.3%

88.8%

92.1%

92.3%

84.3%

   miR-KDE

93.4%

92.8%

93.1%

93.2%

86.3%

with miPred's %SP

96.1%

88.8%

92.5%

92.7%

85.2%

Arthropoda

   miPred

98.8%

89.0%

93.9%

94.2%

88.2%

   miR-KDE

100.0%

92.0%

96.0%

96.2%

92.3%

Viridiplantae

   miPred

98.2%

93.6%

95.9%

96.0%

91.9%

   miR-KDE

98.4%

95.0%

96.7%

96.8%

93.4%

Nematoda

   miPred

97.2%

90.4%

93.8%

94.0%

87.8%

   miR-KDE

97.2%

92.7%

94.9%

95.0%

89.9%

Viruses

   miPred

97.2%

93.1%

95.1%

95.2%

90.4%

   miR-KDE

94.4%

97.2%

95.8%

95.8%

91.7%

with miPred's %SP

98.6%

93.1%

95.8%

95.9%

91.8%

Overall

   miPred

97.3% ± 1.3%

91.0% ± 2.3%

94.1% ± 1.5%

94.3% ± 1.4%

88.5% ± 2.9%

   miR-KDE

96.7% ± 2.7%

93.9% ± 2.1%

95.3% ± 1.4%

95.4% ± 1.4%

90.7% ± 2.8%

with miPred's %SP

98.1% ± 1.5%

92.3% ± 2.2%

95.2% ± 1.6%

95.3% ± 1.6%

90.5% ± 3.3%

  1. The best performance among each dataset is highlighted in bold.