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

Accuracy: 0.429

Confusion matrix
Real	Pred.	#
N	P	86
N	N	226
N	NONE	86
N	NEU	74
NEU	NONE	87
NEU	NEU	66
NEU	N	84
NEU	P	130
NONE	P	54
NONE	N	35
NONE	NONE	45
NONE	NEU	25
P	P	276
P	NEU	56
P	N	51
P	NONE	47

Category evaluation
Cat	P	R	F1
N	0.571	0.479	0.521
NEU	0.299	0.180	0.224
NONE	0.170	0.283	0.212
P	0.505	0.642	0.566

Macro-Precision: 0.386
Macro-Recall: 0.396
Macro-F1: 0.391