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

Accuracy: 0.577

Confusion matrix
Real	Pred.	#
N	N	358
N	NONE	73
N	P	51
N	NEU	9
NEU	N	76
NEU	NONE	39
NEU	P	38
NEU	NEU	11
NONE	N	66
NONE	NONE	119
NONE	P	36
NONE	NEU	3
P	P	224
P	N	57
P	NONE	64
P	NEU	9

Category evaluation
Cat	P	R	F1
N	0.643	0.729	0.683
NEU	0.344	0.067	0.112
NONE	0.403	0.531	0.459
P	0.642	0.633	0.637

Macro-Precision: 0.508
Macro-Recall: 0.490
Macro-F1: 0.499