Evaluation - task1-intertass-CR
2333 tweets in qrel, 1233 tweets in run.

Accuracy: 0.288

Confusion matrix
Real	Pred.	#
N	?	421
N	P	35
N	N	386
N	NONE	56
N	NEU	14
NEU	?	133
NEU	N	90
NEU	P	30
NEU	NONE	29
NEU	NEU	15
NONE	?	223
NONE	NONE	100
NONE	N	91
NONE	P	28
NONE	NEU	5
P	?	323
P	NONE	65
P	N	99
P	P	170
P	NEU	20

Category evaluation
Cat	P	R	F1
N	0.580	0.423	0.489
NEU	0.278	0.051	0.085
NONE	0.400	0.224	0.287
P	0.646	0.251	0.362

Macro-Precision: 0.476
Macro-Recall: 0.237
Macro-F1: 0.317