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

Accuracy: 0.431

Confusion matrix
Real	Pred.	#
N	NONE	97
N	P	61
N	N	232
N	NEU	82
NEU	NONE	118
NEU	NEU	61
NEU	N	95
NEU	P	93
NONE	NONE	48
NONE	NEU	34
NONE	P	46
NONE	N	31
P	P	274
P	NONE	46
P	N	64
P	NEU	46

Category evaluation
Cat	P	R	F1
N	0.550	0.492	0.519
NEU	0.274	0.166	0.207
NONE	0.155	0.302	0.205
P	0.578	0.637	0.606

Macro-Precision: 0.389
Macro-Recall: 0.399
Macro-F1: 0.394