We wonder if we could use Augmented Operation (may be trend forecast) for a car sharing application. The goal is to predict where a user will drop a car after driving it, based on his previous journeys. We understand how to do this when we have a single variable vs time (Volts/t, Processor load/t, bitrate/t, etc), but for this application we have to track a pair of coordinates (longitude and latitude vs time). Is this something that can be achieved with DataMiner?
Unfortunately, we currently don't have multivariate trend prediction in DataMiner. It also seems to me that trend prediction isn't the best approach to achieve what you want to achieve, as it would be unaware of when a journey starts and when it stops. I'd guess your best bet would be to first cut the trend data into distinct journeys, and then train some algorithm based on that data. This algorithm would probably be rather specific for your use-case though, and it is not something that DataMiner is able to do at the moment. You can always make a feature request on https://community.dataminer.services/feature-suggestions, so that we can track how much interest there would be in a feature like this.