Spikes or sudden changes to an individual device can warrant an alert. IoT devices are often subject to movement and weather. They can be affected by times of day or seasons of the year. The fleet of devices could be spread out throughout the world. Trying to get clear insights across the entire fleet can be challenging. Using a machine learning algorithm that incorporates the entire fleet enables us to treat each device separately.
Use cases for Z-Spikes can be a sudden discharge of batteries or a sudden temperature increase. People use Z-Spikes to tell whether something has been jostled or is suddenly vibrating. Z-Spikes can be used on pumps to see whether there is a blockage. Because Z-Spikes do so well across non-homologous environments, they are often a great candidate for edge deployments.
