question Re: Difference between Spark MLlib/ML and H20 in Support Questions

Difference between Spark MLlib/ML and H20

mrizvi — Thu, 19 Oct 2017 23:26:01 GMT

Hi experts,

Just curious to know about the differences between Spark MLlib/ML and H2O in terms of implementation of algorithms, performance and usability and which one is better in what kinds of use-cases?

Thanks a lot in advance.

Re: Difference between Spark MLlib/ML and H20

TimothySpann — Thu, 19 Oct 2017 23:57:07 GMT

You will have to run your algorithms on your cluster with your data to get a reasonable performance analysis.

What language are you looking at?

The Python Spark interface is pretty clean.

http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science.html

H2O has a few more algorithms than Spark MLib.

https://spark.apache.org/docs/latest/ml-classification-regression.html

Re: Difference between Spark MLlib/ML and H20

mrizvi — Fri, 20 Oct 2017 03:40:25 GMT

Thanks @Timothy Spann for your answer. These links are really helpful. I used python for Spark MLlib so will use the same for H2O as well.