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

Accuracy: 0.562

Confusion matrix
Real	Pred.	#
N	N	385
N	NONE	56
N	P	41
N	NEU	9
NEU	N	87
NEU	NONE	32
NEU	P	41
NEU	NEU	4
NONE	NONE	92
NONE	N	89
NONE	P	40
NONE	NEU	3
P	P	212
P	N	83
P	NEU	10
P	NONE	49

Category evaluation
Cat	P	R	F1
N	0.598	0.784	0.678
NEU	0.154	0.024	0.042
NONE	0.402	0.411	0.406
P	0.635	0.599	0.616

Macro-Precision: 0.447
Macro-Recall: 0.455
Macro-F1: 0.451