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

Accuracy: 0.493

Confusion matrix
Real	Pred.	#
N	NONE	131
N	P	29
N	NEU	35
N	N	277
NEU	NONE	170
NEU	N	74
NEU	P	80
NEU	NEU	43
NONE	NONE	81
NONE	NEU	10
NONE	N	22
NONE	P	46
P	P	303
P	NONE	76
P	NEU	29
P	N	22

Category evaluation
Cat	P	R	F1
N	0.701	0.587	0.639
NEU	0.368	0.117	0.178
NONE	0.177	0.509	0.263
P	0.662	0.705	0.682

Macro-Precision: 0.477
Macro-Recall: 0.480
Macro-F1: 0.478