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

Accuracy: 0.363

Confusion matrix
Real	Pred.	#
N	NEU	129
N	P	93
N	N	164
N	NONE	86
NEU	NONE	86
NEU	NEU	100
NEU	N	79
NEU	P	102
NONE	NONE	40
NONE	NEU	38
NONE	N	36
NONE	P	45
P	N	59
P	NONE	70
P	NEU	87
P	P	214

Category evaluation
Cat	P	R	F1
N	0.485	0.347	0.405
NEU	0.282	0.272	0.277
NONE	0.142	0.252	0.181
P	0.471	0.498	0.484

Macro-Precision: 0.345
Macro-Recall: 0.342
Macro-F1: 0.344