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

Accuracy: 0.409

Confusion matrix
Real	Pred.	#
N	NONE	156
N	P	39
N	N	199
N	NEU	78
NEU	NONE	159
NEU	NEU	69
NEU	N	68
NEU	P	71
NONE	P	38
NONE	NONE	73
NONE	NEU	19
NONE	N	29
P	N	36
P	NEU	54
P	NONE	97
P	P	243

Category evaluation
Cat	P	R	F1
N	0.599	0.422	0.495
NEU	0.314	0.188	0.235
NONE	0.151	0.459	0.227
P	0.621	0.565	0.592

Macro-Precision: 0.421
Macro-Recall: 0.408
Macro-F1: 0.415