Session Ready
Exercise

DNN classifier training using Estimators

As you've seen, Estimators can be a good way to easily and quickly create a DNN Classifier model, though this is usually at the expense of optimization and tweaking to maximizing accuracy.

In this exercise, you'll train and evaluate a DNN classifier based on the dataset you used with the Keras API, which aims to predict online customer behavior. exercise To note: the first steps, including feature column selection and classifier selection have been completed for you.

Instructions
100 XP
  • Train your dataset using the dnnclassifier, an input_fun named shopping_input_function, and a training dataset named shopper_train.
  • There's a mistake in the code that evaluates the model. Find and correct it.