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

Accuracy: 0.582

Confusion matrix
Real	Pred.	#
N	N	375
N	NEU	15
N	P	60
N	NONE	41
NEU	N	70
NEU	NONE	33
NEU	P	51
NEU	NEU	10
NONE	NONE	88
NONE	N	69
NONE	NEU	11
NONE	P	56
P	NONE	44
P	N	53
P	P	245
P	NEU	12

Category evaluation
Cat	P	R	F1
N	0.661	0.764	0.709
NEU	0.208	0.061	0.094
NONE	0.427	0.393	0.409
P	0.595	0.692	0.640

Macro-Precision: 0.473
Macro-Recall: 0.477
Macro-F1: 0.475