[
https://issues.apache.org/jira/browse/MAHOUT-140?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Deneche A. Hakim updated MAHOUT-140:
------------------------------------
Status: Patch Available (was: Open)
Work in progress...
A working implementation, I tested it on a two-ubuntu-cluster but more tests
are needed. There is a known issue when the number of maps is high, I'll try to
solve it in the next patch.
The main limitation of this implementation is that each mapper loads a copy of
the data in memory, and because I don't know how to actually share the data
between the mappers of the same slave node, this means that is you launch
simultaneously N maps per cluster node, you'll get N copies of the data in each
node's memory !!!
> In-memory mapreduce Random Forests
> ----------------------------------
>
> Key: MAHOUT-140
> URL: https://issues.apache.org/jira/browse/MAHOUT-140
> Project: Mahout
> Issue Type: New Feature
> Components: Classification
> Affects Versions: 0.2
> Reporter: Deneche A. Hakim
> Priority: Minor
>
> Each mapper is responsible for growing a number of trees with a whole copy of
> the dataset loaded in memory, it uses the reference implementation's code to
> build each tree and estimate the oob error.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
