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

Accuracy: 0.520

Confusion matrix
Real	Pred.	#
N	N	377
N	P	40
N	NONE	45
N	NEU	10
NEU	P	87
NEU	NONE	124
NEU	N	150
NEU	NEU	6
NONE	NONE	61
NONE	N	53
NONE	P	41
NONE	NEU	4
P	P	298
P	N	84
P	NONE	36
P	NEU	12

Category evaluation
Cat	P	R	F1
N	0.568	0.799	0.664
NEU	0.188	0.016	0.030
NONE	0.229	0.384	0.287
P	0.639	0.693	0.665

Macro-Precision: 0.406
Macro-Recall: 0.473
Macro-F1: 0.437