I understand that's a drawback of k-means. I was just wondering if this was something that Spark solved natively. Looks like the answer is no for the time being

I understand that's a drawback of k-means. I was just wondering if this was something that Spark solved natively. Looks like the answer is no for the time being. Thanks for the pointer to LWPR :)