Evaluation - task1-intertass-CR
1233 tweets in qrel, 1233 tweets in run.

Accuracy: 0.561

Confusion matrix
Real	Pred.	#
N	N	330
N	NEU	38
N	NONE	69
N	P	54
NEU	N	65
NEU	NONE	36
NEU	P	45
NEU	NEU	18
NONE	NONE	116
NONE	N	56
NONE	P	46
NONE	NEU	6
P	P	228
P	N	56
P	NONE	53
P	NEU	17

Category evaluation
Cat	P	R	F1
N	0.651	0.672	0.661
NEU	0.228	0.110	0.148
NONE	0.423	0.518	0.466
P	0.611	0.644	0.627

Macro-Precision: 0.478
Macro-Recall: 0.486
Macro-F1: 0.482