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

Accuracy: 0.430

Confusion matrix
Real	Pred.	#
N	NEU	116
N	P	32
N	NONE	127
N	N	197
NEU	NEU	110
NEU	N	66
NEU	NONE	140
NEU	P	51
NONE	NONE	65
NONE	N	24
NONE	NEU	37
NONE	P	33
P	NEU	82
P	NONE	63
P	N	43
P	P	242

Category evaluation
Cat	P	R	F1
N	0.597	0.417	0.491
NEU	0.319	0.300	0.309
NONE	0.165	0.409	0.235
P	0.676	0.563	0.614

Macro-Precision: 0.439
Macro-Recall: 0.422
Macro-F1: 0.430