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

Accuracy: 0.331

Confusion matrix
Real	Pred.	#
N	NONE	228
N	N	183
N	P	33
N	NEU	28
NEU	NONE	225
NEU	NEU	13
NEU	N	89
NEU	P	40
NONE	NONE	100
NONE	N	30
NONE	P	24
NONE	NEU	5
P	N	57
P	P	176
P	NONE	177
P	NEU	20

Category evaluation
Cat	P	R	F1
N	0.510	0.388	0.440
NEU	0.197	0.035	0.060
NONE	0.137	0.629	0.225
P	0.645	0.409	0.501

Macro-Precision: 0.372
Macro-Recall: 0.365
Macro-F1: 0.369