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

Accuracy: 0.488

Confusion matrix
Real	Pred.	#
N	P	31
N	N	276
N	NEU	131
N	NONE	34
NEU	NEU	125
NEU	N	111
NEU	NONE	66
NEU	P	65
NONE	NONE	36
NONE	N	39
NONE	P	40
NONE	NEU	44
P	P	260
P	NEU	114
P	N	40
P	NONE	16

Category evaluation
Cat	P	R	F1
N	0.592	0.585	0.588
NEU	0.302	0.341	0.320
NONE	0.237	0.226	0.232
P	0.657	0.605	0.630

Macro-Precision: 0.447
Macro-Recall: 0.439
Macro-F1: 0.443