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

Accuracy: 0.377

Confusion matrix
Real	Pred.	#
N	N	137
N	NEU	228
N	P	44
N	NONE	63
NEU	NEU	169
NEU	N	62
NEU	NONE	79
NEU	P	57
NONE	NEU	67
NONE	NONE	37
NONE	P	26
NONE	N	29
P	N	58
P	NEU	125
P	P	195
P	NONE	52

Category evaluation
Cat	P	R	F1
N	0.479	0.290	0.361
NEU	0.287	0.460	0.354
NONE	0.160	0.233	0.190
P	0.606	0.453	0.519

Macro-Precision: 0.383
Macro-Recall: 0.359
Macro-F1: 0.371