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

Accuracy: 0.398

Confusion matrix
Real	Pred.	#
N	P	50
N	NONE	149
N	N	231
N	NEU	42
NEU	NEU	39
NEU	NONE	162
NEU	N	89
NEU	P	77
NONE	NONE	74
NONE	N	46
NONE	P	28
NONE	NEU	11
P	N	63
P	P	224
P	NONE	119
P	NEU	24

Category evaluation
Cat	P	R	F1
N	0.538	0.489	0.513
NEU	0.336	0.106	0.161
NONE	0.147	0.465	0.223
P	0.591	0.521	0.554

Macro-Precision: 0.403
Macro-Recall: 0.396
Macro-F1: 0.399