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

Accuracy: 0.369

Confusion matrix
Real	Pred.	#
N	NEU	253
N	P	41
N	NONE	69
N	N	109
NEU	NEU	191
NEU	NONE	77
NEU	N	43
NEU	P	56
NONE	NEU	76
NONE	NONE	37
NONE	P	24
NONE	N	22
P	N	39
P	NEU	148
P	P	190
P	NONE	53

Category evaluation
Cat	P	R	F1
N	0.512	0.231	0.318
NEU	0.286	0.520	0.369
NONE	0.157	0.233	0.187
P	0.611	0.442	0.513

Macro-Precision: 0.391
Macro-Recall: 0.356
Macro-F1: 0.373