Evaluation - task1-intertass-ST4-PE
1428 tweets in qrel, 1428 tweets in run.

Accuracy: 0.481

Confusion matrix
Real	Pred.	#
N	P	112
N	N	290
N	NEU	32
N	NONE	38
NEU	P	148
NEU	N	123
NEU	NONE	60
NEU	NEU	36
NONE	P	76
NONE	N	50
NONE	NONE	28
NONE	NEU	5
P	P	333
P	NEU	29
P	N	56
P	NONE	12

Category evaluation
Cat	P	R	F1
N	0.559	0.614	0.585
NEU	0.353	0.098	0.154
NONE	0.203	0.176	0.189
P	0.498	0.774	0.606

Macro-Precision: 0.403
Macro-Recall: 0.416
Macro-F1: 0.409