Hands-On Deep Learning for IoT

ML techniques for predictive maintenance

Deep learning (DL) techniques can be applied to process the massive amount of IoT data and can be an appealing emerging alternative to classical machine learning algorithms. The idea is that when equipment is given with sensors and networked, a huge amount of sensor data is produced. In a more complex industrial setting, data from the sensor channels is quite noisy and fluctuates over time, but some of the data does not seem to change at all. This is more-or-less true for every industrial setting because the data produced in an IoT setting is a multivariate series of sensor measurements each with its own amount of noise containing many missing or uninformative values.

A key step in predictive maintenance ...