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

Accuracy: 0.299

Confusion matrix
Real	Pred.	#
N	NEU	12
N	P	445
N	N	13
N	NONE	2
NEU	P	343
NEU	N	6
NEU	NEU	16
NEU	NONE	2
NONE	P	149
NONE	NEU	6
NONE	N	3
NONE	NONE	1
P	P	397
P	NEU	25
P	N	5
P	NONE	3

Category evaluation
Cat	P	R	F1
N	0.481	0.028	0.052
NEU	0.271	0.044	0.075
NONE	0.125	0.006	0.012
P	0.298	0.923	0.450

Macro-Precision: 0.294
Macro-Recall: 0.250
Macro-F1: 0.270