Hi, I should forecast the next value of a time series from the series itself and the time series of an external variable influencing it (e.g. room temperature in 1 minute, knowing room temperature and the power being used by the air conditioner), with a model that can be identified/trained using many weeks of signals.
I already have the two series cleansed, synched and without holes.
Any suggestion as the standard or most common approach?
Thanks!!
Hi,
Welcome to the KNIME Forum! You can try applying nodes from the KNIME Autoregressive integrated moving average (ARIMA) extension. If you need more firepower, try an LSTM like in this example: Time Series Prediction. We also have nodes for removing seasonality before building a model: Remove Seasonality.
Kind regards
Alexander
3 Likes
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.