Simple Regression Using TensorFlow
This section will explain, step by step, how to successfully tackle a regression problem. You will learn how to take a preliminary look at the dataset to understand its most important properties, as well as how to prepare it to be used during training, validation, and inference. Then, a deep neural network will be built from a clean sheet using TensorFlow via the Keras API. This model will then be trained and its performance will be evaluated.
In a regression problem, the aim is to predict the output of a continuous value, such as a price or a probability. In this exercise, the classic Auto MPG dataset will be used and a deep neural network will be trained on it to accurately predict car fuel efficiency, using no more than the following seven features: Cylinders, Displacement, Horsepower, Weight, Acceleration, Model Year, and Origin.
The dataset can be thought of as a table with eight columns (seven features, plus one target value) and as...