Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/3022#issuecomment-67580723
Here are some results I got using the text8-100 dataset. It's just a local
test (1 worker), but we can do larger-scale tests in the future.
numInstances | k | numFeatures |
time(sec) | avg kmeansCost*
------------- | ------- | ------- | ------- |
-------
170053 | 2 | 10 | 46.53373526 |
0.772718978
170053 | 4 | 10 | 65.32307584 |
0.720132151
170053 | 16 | 10 | 225.6346969 |
0.632446005
170053 | 64 | 10 | 894.4889346 |
0.525024814
170053 | 2 | 100 | 39.96043946 |
76.93881368
170053 | 4 | 100 | 65.9859325 |
76.93881368
170053 | 16 | 100 | 224.1422487 |
76.93881368
170053 | 64 | 100 | 875.0000443 |
76.93881368
\* avg squared L2 distance
I'll make one more pass, but I think this is basically ready.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
