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

Accuracy: 0.522

Confusion matrix
Real	Pred.	#
N	N	274
N	NONE	130
N	NEU	51
N	P	36
NEU	N	43
NEU	NONE	57
NEU	P	41
NEU	NEU	23
NONE	NONE	134
NONE	P	39
NONE	NEU	16
NONE	N	35
P	NONE	83
P	N	33
P	P	213
P	NEU	25

Category evaluation
Cat	P	R	F1
N	0.712	0.558	0.626
NEU	0.200	0.140	0.165
NONE	0.332	0.598	0.427
P	0.647	0.602	0.624

Macro-Precision: 0.473
Macro-Recall: 0.475
Macro-F1: 0.474