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

Accuracy: 0.449

Confusion matrix
Real	Pred.	#
N	NEU	82
N	P	53
N	NONE	102
N	N	235
NEU	NONE	117
NEU	NEU	77
NEU	N	79
NEU	P	94
NONE	P	52
NONE	NONE	49
NONE	N	34
NONE	NEU	24
P	P	280
P	NEU	57
P	NONE	46
P	N	47

Category evaluation
Cat	P	R	F1
N	0.595	0.498	0.542
NEU	0.321	0.210	0.254
NONE	0.156	0.308	0.207
P	0.585	0.651	0.616

Macro-Precision: 0.414
Macro-Recall: 0.417
Macro-F1: 0.415