Hi ,
Have anyone ideas about how to choose cells with missing values and replace them with a predicted value?
I’ve tried to loop over all the columns containing missing values and with random forest to get the predict values of each column…I’m wondering to how can I use the predicted value in the columns to replace the missing one…Thanks.
The training data contains all the rows without missing values and the test data contains the rows with missing values… I get for every column a column with predicted value…I want to only take the cells containing missing values and replace those with a predicted one I get…
Best Regards,
N.
Hey @Northern,
just ask the obvious question: Did you try using the missing value node?
Is the “predicted value” mentioned by you a static value at all time or does it changes from run to run or during the run of a workflow?
From my experience, with a more detail explanation and some example data, the community can help you much easier.
Kind regards,
Patrick
Hi Patrick,
I’ve tried using a random forest to learn and predict the missing values in a loop…every time would one column with missing values selected (here I used a column loop start node to choose and run for each column)
The training data contains all the rows without missing values and the test data contains the rows with missing values… I get for every column a column with predicted value, but what I’m wondering is only take the cells that really contain missing values and replace these cells…
I thought that the Column Merger Node would help but I don’t know how can I use it for more than one columns…
Best Regards,
N.
