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

Accuracy: 0.460

Confusion matrix
Real	Pred.	#
N	N	285
N	NEU	108
N	P	98
NEU	N	57
NEU	NEU	43
NEU	P	64
NONE	N	52
NONE	P	88
NONE	NEU	83
NONE	NONE	1
P	NEU	52
P	P	238
P	N	63
P	NONE	1

Category evaluation
Cat	P	R	F1
N	0.624	0.580	0.601
NEU	0.150	0.262	0.191
NONE	0.500	0.004	0.009
P	0.488	0.672	0.565

Macro-Precision: 0.440
Macro-Recall: 0.380
Macro-F1: 0.408