Mastering Microsoft Dynamics 365 Business Central

Step 4 – Getting insights into how ML works

When you get predictions from an ML web service, it's always interesting to understand why the model gives some results or makes those decisions. As the ML Prediction API uses tree-based ML algorithms, we can see the decision tree:

It's important to understand that the decision tree is an artifact of the trained model, not of the predictions made by the trained model.

Open the Train Rest. Forecast ML codeunit and find the DownloatPlotOfTheModel function. Let's see how it works:

1. Connect to your published ML web service.
2. Get the .pdf file with the decision tree calling the PlotModel function. You ...