AutoML using the AML Python SDK

Now that you have seen how to leverage AutoML inside AMLS, we will explore creating a model leveraging the AML Python SDK.

Earlier in the chapter, in Introduction to AutoML, we cloned our sample notebooks to leverage this material. For this chapter, note that there is a single notebook titled Chapter5_Titanic_AutoML. The initial code should look familiar, as the preparation of the dataset has not changed. However, when we run the experiment, we will now be leveraging an AutoML experiment.

In order to run an AutoML experiment, we will need to import AutoMLConfig as shown in line 8 of the following figure.

Here’s AutoMLConfig being imported with the Python SDK:

Figure 5.17 – Importing AutoMLConfig

As part of an AutoML experiment, we are required to specify a training dataset, but we are not required to specify a validation or testing dataset. However, we can specify the training and validation datasets...