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

Accuracy: 0.528

Confusion matrix
Real	Pred.	#
N	N	308
N	P	34
N	NONE	115
N	NEU	15
NEU	P	97
NEU	NONE	149
NEU	N	88
NEU	NEU	33
NONE	NONE	73
NONE	N	29
NONE	P	51
NONE	NEU	6
P	P	340
P	NEU	24
P	NONE	38
P	N	28

Category evaluation
Cat	P	R	F1
N	0.680	0.653	0.666
NEU	0.423	0.090	0.148
NONE	0.195	0.459	0.273
P	0.651	0.791	0.714

Macro-Precision: 0.487
Macro-Recall: 0.498
Macro-F1: 0.493