[
https://issues.apache.org/jira/browse/MADLIB-976?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Frank McQuillan updated MADLIB-976:
-----------------------------------
Description: When running Random Forest training function on a modest data
set it took a long time - much longer than expected or desired. The training
data set has around 8000 rows and 400 features. Several models on similar data
all took around 40,000 seconds to run. Each time we used 100 trees and 14
random features selected. (was: When running Random Forest training function
on a modest data set it took a long time - much longer than expected or
desired. The training data set has around 8000 rows and 400 features. Several
models on similar data all took around 40,000 seconds to run. Each time we
used 100 trees and 14 random features selected. Dataset shared offline.)
> Random Forest - extremely long training time
> --------------------------------------------
>
> Key: MADLIB-976
> URL: https://issues.apache.org/jira/browse/MADLIB-976
> Project: Apache MADlib
> Issue Type: Bug
> Components: Module: Random Forest
> Reporter: Rashmi Raghu
> Priority: Major
> Fix For: v2.0
>
>
> When running Random Forest training function on a modest data set it took a
> long time - much longer than expected or desired. The training data set has
> around 8000 rows and 400 features. Several models on similar data all took
> around 40,000 seconds to run. Each time we used 100 trees and 14 random
> features selected.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
