)
It can be competitive for high dimensional datasets. 
cons
Generalization performance may worse than more sophisticated learning methods.