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

Accuracy: 0.368

Confusion matrix
Real	Pred.	#
N	NONE	149
N	NEU	101
N	P	48
N	N	174
NEU	NONE	137
NEU	N	80
NEU	NEU	85
NEU	P	65
NONE	P	26
NONE	NONE	61
NONE	N	45
NONE	NEU	27
P	N	43
P	P	206
P	NONE	116
P	NEU	65

Category evaluation
Cat	P	R	F1
N	0.509	0.369	0.428
NEU	0.306	0.232	0.264
NONE	0.132	0.384	0.196
P	0.597	0.479	0.532

Macro-Precision: 0.386
Macro-Recall: 0.366
Macro-F1: 0.376