Feature importance IN random forest

[pseudo oduesp]

Hi,
 i use pyspark 1.5.0
can i  ask you how i can get feature imprtance for a randomforest
algorithme in pyspark and please give me example
thanks for advance.