An accelerometer measures the x, y, and z component of accelerations, as shown in the following diagram:
This property of the accelerometer enables it to be placed in a wearable device, such as a cell phone mounted on a person's wrist with a wrist band, smartwatch, or even in a shoe, to measure the XYZ component of acceleration. In this section, we will learn how accelerometer data can be analyzed using neural networks to identify human activity. We will develop a machine learning model with TensorFlow. This is the only section of this book that discusses how to use raw data without an image and how to pass that to a neural network to develop a model and draw inference from it.
Human activity recognition involves classifying different types of activity based on accelerometer data. The challenge here is to correlate accelerometer data...
