MIT Press

Table 9:

Test accuracy for tags that were observed 0, <100, and ≥ 100 times in the 1000 sentence training sets. Tok is the number of tokens with the specified count, Tag is the accuracy using a whole-tag classifier, Seq is the accuracy using a sequence decoder.

	count=0			count<100			count≥100
Lang	Tok	Tag	Seq	Tok	Tag	Seq	Tok	Tag	Seq
SV	12	0.0	8.33	844	81.28	82.82	19521	94.49	94.65
BG	4	0.0	0.0	910	81.32	83.41	14810	96.62	97.37
HU	108	0.0	20.37	2333	53.54	59.24	8007	78.24	80.67
PT	3	0.0	0.0	207	63.29	67.63	9991	93.04	92.25

	count=0			count<100			count≥100
Lang	Tok	Tag	Seq	Tok	Tag	Seq	Tok	Tag	Seq
SV	12	0.0	8.33	844	81.28	82.82	19521	94.49	94.65
BG	4	0.0	0.0	910	81.32	83.41	14810	96.62	97.37
HU	108	0.0	20.37	2333	53.54	59.24	8007	78.24	80.67
PT	3	0.0	0.0	207	63.29	67.63	9991	93.04	92.25