Skip to main content
. 2008 Oct 5;78(5):308–320. doi: 10.1016/j.ijmedinf.2008.08.004

Table 4.

Performance comparison for MIM and Google Translation

Syndrome TP + FN PPV Sensitivity Specificity F F2
Mutual Information-based Mapping (MIM)
 GI 592 0.97*** 0.97*** 0.98*** 0.97*** 0.97***
 RASH 45 0.87** 0.77 0.99** 0.82*** 0.80***
 RESP 331 0.89*** 0.96*** 0.97** 0.93*** 0.94***
 URESP 132 0.86*** 0.91** 0.98*** 0.88*** 0.89***
 LRESP 272 0.93 0.98*** 0.98 0.95*** 0.96***
 FEVER 413 0.99** 0.96 0.99** 0.97 0.97



Google Translation
 GI 592 0.91 0.90 0.96 0.91 0.91
 RASH 45 0.76 0.73 0.99 0.75 0.74
 RESP 331 0.84 0.83 0.96 0.83 0.83
 URESP 132 0.70 0.83 0.97 0.76 0.78
 LRESP 272 0.96** 0.80 0.99*** 0.87 0.84
 FEVER 413 0.98 0.96 0.99 0.97 0.97

Statistical test is based on 3000 bootstrappings.

*p-value <0.1; **p-value <0.05; ***p-value <0.01.