Skip to main content

Table 6 Comparison on AImed. (P)recision, (R)ecall, (F)-score and AUC results for methods evaluated on AImed with the correct cross-validation methodology. Note that the best performing method, introduced by Miwa et al. [34], also utilizes the all-paths graph kernel.

From: All-paths graph kernel for protein-protein interaction extraction with evaluation of cross-corpus learning

  P R F AUC
Miwa et al. [34] - - 63.5% 87.9%
Miyao et al. [35] 54.9% 65.5% 59.5% -
Giuliano et al. [28] 60.9% 57.2% 59.0% -
All-paths graph kernel 52.9% 61.8% 56.4% 84.8%
Sætre et al. [32] 64.3% 44.1% 52.0% -
Mitsumori et al. [39] 54.2% 42.6% 47.7% -
Van Landeghem et al. [29] 49% 44% 46% -
Yakushiji et al. [40] 33.7% 33.1% 33.4% -
\