Realtime forecasting

With TIM’s RTInstantML, producing forecasts instantly from data is affordable. TIM accomplishes this by looking at the latest available data and detecting a corresponding forecasting routine.

There is no need to set up a data availability scheme for the current situation – this is defined by the way the data is organized and TIM will recognize this automatically.

The user only needs to provide TIM with the number of steps ahead for which the forecast should be calculated. TIM will then generate a model and produce a forecast in one single API call.

The model is discarded as it will not be needed anymore: the next time a forecast is required the entire process will be repeated. This is inherently different to AutoML, where a model is built once and is then reused many times for calculating forecasts.