Table 3.

Results on the 2014 SemEval data set.

Features	Precision	Recall	F1-score
Baseline features	0.782	0.727	0.754
+RawEmb (ranking)	0.775	0.758	0.767
+BinEmb (ranking)	0.781	0.774	0.778
+ProtoEmb (ranking)	0.784	0.748	0.766
+RawEmb (word2vec)	0.778	0.750	0.764
+BinEmb (word2vec)	0.779	0.764	0.771
+ProtoEmb (word2vec)	0.789	0.752	0.770
+BrownCluster	0.778	0.756	0.767
+BrownCluster+BinEmb (ranking)	0.783	0.780	0.781