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

Accuracy: 0.563

Confusion matrix
Real	Pred.	#
N	N	375
N	P	59
N	NONE	47
N	NEU	10
NEU	N	76
NEU	NONE	24
NEU	P	48
NEU	NEU	16
NONE	N	82
NONE	NONE	84
NONE	P	42
NONE	NEU	16
P	P	219
P	N	73
P	NONE	41
P	NEU	21

Category evaluation
Cat	P	R	F1
N	0.619	0.764	0.684
NEU	0.254	0.098	0.141
NONE	0.429	0.375	0.400
P	0.595	0.619	0.607

Macro-Precision: 0.474
Macro-Recall: 0.464
Macro-F1: 0.469