[
https://issues.apache.org/jira/browse/MADLIB-928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15229279#comment-15229279
]
ASF GitHub Bot commented on MADLIB-928:
---------------------------------------
GitHub user orhankislal opened a pull request:
https://github.com/apache/incubator-madlib/pull/39
Random Forest: Fix filtered feature related bug
JIRA: MADLIB-928
Additional Author: Nandish Jayaram <[email protected]>
Random forest filters out a feature if it has the same value for every
row. If grouping is enabled, this filer is applied per group. However
_get_bins_grps function did not consider a case where different groups
have different feature sets. The commit fixes this issue, adds a test
for install-check to replicate the use case and fixes a typo in one of
the related functions.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/orhankislal/incubator-madlib
bugfix/random_forest3
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-madlib/pull/39.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #39
----
commit 1b75152b1245f11b6aa449b4e84189468b40e92d
Author: Orhan Kislal <[email protected]>
Date: 2016-04-06T21:07:26Z
Random Forest: Fix filtered feature related bug
JIRA: MADLIB-928
Additional Author: Nandish Jayaram <[email protected]>
Random forest filters out a feature if it has the same value for every
row. If grouping is enabled, this filer is applied per group. However
_get_bins_grps function did not consider a case where different groups
have different feature sets. The commit fixes this issue, adds a test
for install-check to replicate the use case and fixes a typo in one of
the related functions.
----
> error in Random Forest
> -----------------------
>
> Key: MADLIB-928
> URL: https://issues.apache.org/jira/browse/MADLIB-928
> Project: Apache MADlib
> Issue Type: Bug
> Components: Module: Random Forest
> Reporter: Tetsuo Kobayashi
> Assignee: Orhan Kislal
> Priority: Minor
> Fix For: v1.9.1
>
>
> I am currently having an error with the MADlib Random Forest function in
> MADlib 1.8.0. Below is the code I tried.
> SELECT madlib.forest_train('data_random_forest', -- input table name
> 'rf_output', -- output table name
> 'id', -- id column
> 'duration', -- dependent variable
> '*', -- list of features
> 'start_time, end_time, date_only, loc_code,
> amedas_code', -- exclude columns
> 'linkid' -- grouping column
> ,10::integer -- # of trees
> ,2::integer, -- # of random features
> TRUE::boolean, -- importance
> 1, -- # of permutations
> 10, -- max_tree_depth
> 8, -- min_split
> 3, -- min_bucket
> 10 -- number of splits per continuous variable
> );
> I got an error which tells me something is wrong with the feature
> "holiday_flg) which is a dummy variable. When I excluded "holiday_flg," the
> error message tells me that the "KeyError: 'sun_flg' which is another dummy
> variable. I got another dummy variable in "KeyError:" when I excluded
> "sun_flg." I can tell one of the dummy variables in the model is picked up in
> the error message when I exclude one.
> I was told by Rahul Iyer that this is most probably because at least 1 group
> has only 1 level for (some) categorical variables. This can happen and I saw
> some groups have identical value for all records.
> I am using MADlib 1.8.0 on GPDB 4.3.6.1. The OS is CentOS.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
