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

Accuracy: 0.319

Confusion matrix
Real	Pred.	#
N	P	26
N	NONE	249
N	N	194
N	NEU	3
NEU	NONE	277
NEU	P	33
NEU	N	52
NEU	NEU	5
NONE	NONE	120
NONE	N	19
NONE	NEU	2
NONE	P	18
P	NONE	242
P	P	137
P	N	50
P	NEU	1

Category evaluation
Cat	P	R	F1
N	0.616	0.411	0.493
NEU	0.455	0.014	0.026
NONE	0.135	0.755	0.229
P	0.640	0.319	0.425

Macro-Precision: 0.461
Macro-Recall: 0.374
Macro-F1: 0.413