Evaluation - task1-intertass-PE
1428 tweets in qrel, 1428 tweets in run.

Accuracy: 0.463

Confusion matrix
Real	Pred.	#
N	NEU	39
N	P	48
N	NONE	105
N	N	280
NEU	NONE	138
NEU	NEU	51
NEU	N	105
NEU	P	73
NONE	P	51
NONE	NONE	64
NONE	N	32
NONE	NEU	12
P	P	266
P	NEU	33
P	N	65
P	NONE	66

Category evaluation
Cat	P	R	F1
N	0.581	0.593	0.587
NEU	0.378	0.139	0.203
NONE	0.172	0.403	0.241
P	0.607	0.619	0.613

Macro-Precision: 0.434
Macro-Recall: 0.438
Macro-F1: 0.436