在Azure Machine Learning Studio中无需代码合并分类列值的可行方案

Absolutely, you can pull this off without writing any SQL, R, or Python—Azure ML Studio has built-in visual modules that handle exactly this kind of category mapping. Here's how to do it step by step:

This is the go-to module for remapping categorical values in Azure ML Studio, and it’s perfect for your use case:

1. Add the Map Values module to your experiment
   You’ll find it under the Data Transformation > Manipulation category in the module palette on the left side of the studio. Drag and drop it onto your canvas.

2. Connect your dataset
   Link the output of your existing dataset (the one with the 'yes'/'no'/'maybe' column) to the input port of the Map Values module.

3. Configure the value mappings
   Open the module’s settings panel (click the module, then the "Settings" tab on the right):

   • Under Column to transform, select the specific column you want to modify.
   • Click the Launch mapper button to open the value mapping editor.
   • In the editor table:
     • Locate the row with the original value maybe, then change its corresponding Target Value to no.
     • Leave yes mapped to yes and no mapped to no (these should stay as-is unless you need further changes).
   • Click OK to save your mapping rules.

4. Run the experiment
   Once you run the experiment, the output dataset from the Map Values module will have your target column updated: all 'maybe' entries will now be 'no', leaving you with just two categories: 'yes' and 'no'.

Quick Tip #

If your target column isn’t already marked as a categorical type, use the Edit Metadata module first to set its type to Categorical. This ensures the Map Values module correctly identifies and processes all the category values without issues.

内容的提问来源于stack exchange，提问作者Danrex