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

Accuracy: 0.319

Confusion matrix
Real	Pred.	#
N	NONE	323
N	P	24
N	N	123
N	NEU	2
NEU	NONE	278
NEU	N	43
NEU	P	44
NEU	NEU	2
NONE	NONE	123
NONE	P	25
NONE	N	11
P	N	33
P	NONE	190
P	P	207

Category evaluation
Cat	P	R	F1
N	0.586	0.261	0.361
NEU	0.500	0.005	0.011
NONE	0.135	0.774	0.229
P	0.690	0.481	0.567

Macro-Precision: 0.478
Macro-Recall: 0.380
Macro-F1: 0.423