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

Accuracy: 0.395

Confusion matrix
Real	Pred.	#
N	N	226
N	NONE	185
N	P	32
N	NEU	29
NEU	NONE	242
NEU	P	45
NEU	N	61
NEU	NEU	19
NONE	NONE	107
NONE	N	21
NONE	P	26
NONE	NEU	5
P	NONE	161
P	P	212
P	N	35
P	NEU	22

Category evaluation
Cat	P	R	F1
N	0.659	0.479	0.555
NEU	0.253	0.052	0.086
NONE	0.154	0.673	0.251
P	0.673	0.493	0.569

Macro-Precision: 0.435
Macro-Recall: 0.424
Macro-F1: 0.429