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

Accuracy: 0.201

Confusion matrix
Real	Pred.	#
N	NEU	81
N	NONE	252
N	N	82
N	P	57
NEU	NONE	200
NEU	NEU	61
NEU	N	57
NEU	P	49
NONE	NONE	92
NONE	P	17
NONE	N	28
NONE	NEU	22
P	P	52
P	NONE	244
P	NEU	66
P	N	68

Category evaluation
Cat	P	R	F1
N	0.349	0.174	0.232
NEU	0.265	0.166	0.204
NONE	0.117	0.579	0.194
P	0.297	0.121	0.172

Macro-Precision: 0.257
Macro-Recall: 0.260
Macro-F1: 0.258