A detailed error analysis of 13 kernel methods for protein-protein interaction extraction

BMC Bioinformatics

Table 5 Classification results on the 521 ND pairs with CL evaluation

Kernel	r	TN	e	TN/e	TN/#ND
SST	26.9	288	381	0.76	0.55
edit	22.5	279	404	0.69	0.54
ST	29.2	231	369	0.63	0.44
APG	26.9	207	381	0.54	0.40
SL	29.9	177	365	0.48	0.34
lexical	24.5	170	393	0.43	0.33
cosine	26.6	157	382	0.41	0.30
syntactic	26.9	155	381	0.41	0.30
SpT	42.1	142	302	0.47	0.27
combined	26.8	132	381	0.35	0.25
shallow	28.6	127	372	0.34	0.24
kBSPS	37.1	120	328	0.37	0.23

Classification results on the 521 ND pairs with CL evaluation (in decreasing order according to the number of successfully classified pairs). Ratio (r) refers to the distribution of positive classes predicted by the kernel measured across the 5 corpora; TN is the number of correctly classified ND pairs; e is 521·(1−r), the expected number of negative class predictions projected onto the 521 ND pairs.

ISSN: 1471-2105