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

Accuracy: 0.424

Confusion matrix
Real	Pred.	#
N	NONE	142
N	P	55
N	N	255
N	NEU	20
NEU	NONE	147
NEU	N	120
NEU	NEU	21
NEU	P	79
NONE	NONE	79
NONE	P	37
NONE	N	37
NONE	NEU	6
P	P	250
P	NONE	82
P	NEU	25
P	N	73

Category evaluation
Cat	P	R	F1
N	0.526	0.540	0.533
NEU	0.292	0.057	0.096
NONE	0.176	0.497	0.259
P	0.594	0.581	0.588

Macro-Precision: 0.397
Macro-Recall: 0.419
Macro-F1: 0.408