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

Accuracy: 0.494

Confusion matrix
Real	Pred.	#
N	P	59
N	NONE	90
N	N	315
N	NEU	8
NEU	NONE	156
NEU	P	78
NEU	N	109
NEU	NEU	24
NONE	NONE	74
NONE	N	42
NONE	NEU	6
NONE	P	37
P	P	292
P	N	67
P	NONE	60
P	NEU	11

Category evaluation
Cat	P	R	F1
N	0.591	0.667	0.627
NEU	0.490	0.065	0.115
NONE	0.195	0.465	0.275
P	0.627	0.679	0.652

Macro-Precision: 0.476
Macro-Recall: 0.469
Macro-F1: 0.472