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

Accuracy: 0.537

Confusion matrix
Real	Pred.	#
N	NEU	115
N	N	311
N	P	32
N	NONE	33
NEU	N	58
NEU	NEU	64
NEU	P	27
NEU	NONE	15
NONE	N	60
NONE	NONE	81
NONE	P	40
NONE	NEU	43
P	P	206
P	N	42
P	NEU	78
P	NONE	28

Category evaluation
Cat	P	R	F1
N	0.660	0.633	0.647
NEU	0.213	0.390	0.276
NONE	0.516	0.362	0.425
P	0.675	0.582	0.625

Macro-Precision: 0.516
Macro-Recall: 0.492
Macro-F1: 0.504