[
https://issues.apache.org/jira/browse/MADLIB-1323?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16824373#comment-16824373
]
Nandish Jayaram commented on MADLIB-1323:
-----------------------------------------
MLP already has the code to create these columns (albeit with the wrong column
name currently). We refactored the code out to function
{{create_cols_from_array_sql_string}} in {{utilities.py_in}} to generate the
necessary query for creation of these columns. This fix to this JIRA is to use
that function and refactoring the code in MLP. Refer to
https://github.com/apache/madlib/blob/master/src/ports/postgres/modules/deep_learning/madlib_keras_predict.py_in#L111
for how it is used.
> MLP - fix column name in output table
> -------------------------------------
>
> Key: MADLIB-1323
> URL: https://issues.apache.org/jira/browse/MADLIB-1323
> Project: Apache MADlib
> Issue Type: Bug
> Components: Module: Neural Networks
> Reporter: Frank McQuillan
> Priority: Minor
> Fix For: v1.16
>
>
> http://madlib.apache.org/docs/latest/group__grp__nn.html#example
> {code}
> output_table
> TEXT. Name of the table where output predictions are written. If this table
> name is already in use, an error is returned. Table contains:
> id Gives the 'id' for each prediction, corresponding to each row from the
> data_table.
> estimated_COL_NAME (For pred_type='response') The estimated class for
> classification or value for regression, where COL_NAME is the name of the
> column to be predicted from training data.
> prob_CLASS
> (For pred_type='prob' for classification) The probability of a given class
> CLASS as given by softmax. There will be one column for each class in the
> training data.
> {code}
> But the code outputs `estimated_prob_CLASS`, e.g.
> {code}
> estimated_prob_Iris_setosa | estimated_prob_Iris_versicolor
> {code}
> so we should fix this so it outputs `prob_CLASS`, e.g.
> {code}
> prob_Iris_setosa | prob_Iris_versicolor
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
