Hi All,

  I am running k-means clustering algorithm. Now, when I am running the
algorithm as -

val conf = new SparkConf
val sc = new SparkContext(conf)
.
.
val kmeans = new KMeans()
val model = kmeans.run(RDD[Vector])
.
.
.
The 'kmeans' object gets created on driver. Now does *kmeans.run() *get
executed on each partition of the rdd in distributed fashion or else does
the entire RDD is brought to driver and then gets executed at the driver on
the entire RDD ??

Thanks,
Padma Ch

Reply via email to