Evaluation - task1-intertass-CR
1233 tweets in qrel, 1233 tweets in run.

Accuracy: 0.567

Confusion matrix
Real	Pred.	#
N	N	313
N	P	61
N	NONE	97
N	NEU	20
NEU	N	55
NEU	NONE	54
NEU	P	41
NEU	NEU	14
NONE	N	51
NONE	NONE	120
NONE	P	50
NONE	NEU	3
P	P	252
P	N	40
P	NONE	51
P	NEU	11

Category evaluation
Cat	P	R	F1
N	0.682	0.637	0.659
NEU	0.292	0.085	0.132
NONE	0.373	0.536	0.440
P	0.624	0.712	0.665

Macro-Precision: 0.493
Macro-Recall: 0.493
Macro-F1: 0.493