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

Accuracy: 0.433

Confusion matrix
Real	Pred.	#
N	NONE	115
N	P	82
N	N	234
N	NEU	41
NEU	NONE	118
NEU	N	91
NEU	NEU	48
NEU	P	110
NONE	NONE	56
NONE	P	57
NONE	N	33
NONE	NEU	13
P	NONE	74
P	N	53
P	P	280
P	NEU	23

Category evaluation
Cat	P	R	F1
N	0.569	0.496	0.530
NEU	0.384	0.131	0.195
NONE	0.154	0.352	0.215
P	0.529	0.651	0.584

Macro-Precision: 0.409
Macro-Recall: 0.407
Macro-F1: 0.408