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

Accuracy: 0.432

Confusion matrix
Real	Pred.	#
N	NONE	82
N	P	82
N	N	247
N	NEU	61
NEU	P	106
NEU	NEU	55
NEU	N	109
NEU	NONE	97
NONE	NONE	41
NONE	P	45
NONE	N	53
NONE	NEU	20
P	N	81
P	NONE	46
P	P	274
P	NEU	29

Category evaluation
Cat	P	R	F1
N	0.504	0.523	0.514
NEU	0.333	0.150	0.207
NONE	0.154	0.258	0.193
P	0.540	0.637	0.585

Macro-Precision: 0.383
Macro-Recall: 0.392
Macro-F1: 0.387