Evaluation - task1-intertass-ST4-CR
1233 tweets in qrel, 1233 tweets in run.

Accuracy: 0.569

Confusion matrix
Real	Pred.	#
N	N	413
N	NONE	15
N	P	60
N	NEU	3
NEU	N	112
NEU	P	40
NEU	NONE	7
NEU	NEU	5
NONE	N	117
NONE	NONE	46
NONE	P	60
NONE	NEU	1
P	P	237
P	N	101
P	NONE	11
P	NEU	5

Category evaluation
Cat	P	R	F1
N	0.556	0.841	0.669
NEU	0.357	0.030	0.056
NONE	0.582	0.205	0.304
P	0.597	0.669	0.631

Macro-Precision: 0.523
Macro-Recall: 0.437
Macro-F1: 0.476