Short Time Prediction

Version: 7.2.0

The Prediction Engine uses machine learning to predict the trend in an iSCAN channel and upload the result back to iSCAN.

It will connect to iSCAN, download the data for the provided input_channels and output_channel, use machine learning to reconstruct the output channel, and upload the results to the given prediction_channel.

In order for the prediction to be accurate, there should be a relation between the input columns and the output column, otherwise the machine learning model will not provide an accurate prediction.

For more technical documentation see the Docs

For support you can reach the PI Team at pit@iesve.com

e.g. https://iscan.iesve.com/building-details/{Project}/{Building}.

You can create a token on the page https://iscan.iesve.com/project-tokens/{Project}.

(optional) Comma-separated list of channels that can influence the channel to be predicted.
Can be Names, IDs, or a combination of them.
e.g. Dry Bulb Temperature, Wet Bulb Temperature, SC00001
Can be left empty if there's no meaningful relationship. In this case only the historical data for the output channel will be used.

Name of ID of the channel that needs to be predicted.
e.g. DO10_kitchen||temperature or SC000014

Name of the channel where the prediction will be uploaded.
e.g. DO10_kitchen||temperature__prediction

How long should the prediction stretch.

Please note that the Prediction Engine can take up to several minutes to successfully run.