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

Accuracy: 0.424

Confusion matrix
Real	Pred.	#
N	NEU	78
N	P	48
N	NONE	141
N	N	205
NEU	NONE	184
NEU	N	53
NEU	P	80
NEU	NEU	50
NONE	NONE	84
NONE	N	19
NONE	P	44
NONE	NEU	12
P	P	267
P	NEU	39
P	NONE	87
P	N	37

Category evaluation
Cat	P	R	F1
N	0.653	0.434	0.522
NEU	0.279	0.136	0.183
NONE	0.169	0.528	0.256
P	0.608	0.621	0.614

Macro-Precision: 0.427
Macro-Recall: 0.430
Macro-F1: 0.429