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

Accuracy: 0.447

Confusion matrix
Real	Pred.	#
N	NONE	150
N	P	36
N	N	257
N	NEU	29
NEU	NONE	193
NEU	NEU	29
NEU	P	60
NEU	N	85
NONE	NONE	78
NONE	N	41
NONE	P	31
NONE	NEU	9
P	P	274
P	NONE	91
P	N	42
P	NEU	23

Category evaluation
Cat	P	R	F1
N	0.605	0.544	0.573
NEU	0.322	0.079	0.127
NONE	0.152	0.491	0.232
P	0.683	0.637	0.659

Macro-Precision: 0.441
Macro-Recall: 0.438
Macro-F1: 0.439