Search code examples
azuremachine-learningdesignerazure-machine-learning-service

How can I select a column product of a math operation in Azure Machine Learning Designer?


I have created a pipeline in Azure Machine Learning that includes a Math Operation (natural logarithm of a column named charges). The next pill to the Math Operation is Select Column in Dataset. Since the pipeline has not ben submitted and run I cannot access the column ln(charges) in the pill Select Column in Dataset. My problem is that if I submit it I am able to run it and see the results in the pipeline once completed, but I have found no way of accessing those results (and thus the ln(charges) column in Designer.

Pipeline Job after submitting and running

Pipeline in designer after submitting and running the job

UPDATE:

I have found a workaround. Still in designer the column ln(charges) is not selectable but if I manually enter Ln(charges) in the select column fields it works.


Solution

  • The following is the procedure of the math operation in Azure ML designer to select the column to be implemented. The following procedure will help to give the column name as well as we can also give the index number of the column. This answer contains both the procedures.

    enter image description here

    We can click on edit column.

    enter image description here

    enter image description here

    Based on the dataset which the experiment was running, both are options are mentioned in the above screen. We can choose either of the options.

    enter image description here

    To access the data, right click and go to access data and click on result_dataset

    enter image description here

    The following page will open and click on any file mentioned in the box

    enter image description here

    Click on download and open in the editor according to your wish.

    enter image description here

    It looks like the above result screen. The below screens are the designer created.

    enter image description here

    enter image description here

    enter image description here

    To check the final model result. Go to evaluate model and get the results in visualization manner.