General settings
The time variable must be a column with the meaning Date.
Single
Multiple

Time step parameters

Business days are from Monday to Friday.
Timestamps are set to the next {{getWeekDayName(mlTaskDesign.timestepParams.endOfWeekDay)}}.
Timestamps in February will be set to the last day of the month instead.
of
Timestamps in February will be set to the last day of the month instead.
of
Timestamps in February will be set to the last day of the month instead.
of
Timestamps in February will be set to the last day of the month instead.

Forecasting parameters

{{ prettyTimeSteps(uiState.predictionLength * uiState.timestepParams.numberOfTimeunits, uiState.timestepParams.timeunit) }}
{{ prettyTimeSteps(mlTaskDesign.evaluationParams.gapSize * uiState.timestepParams.numberOfTimeunits, uiState.timestepParams.timeunit) }}

Quantiles

For statistical and deep learning models, forecasts quantiles give lower and upper bounds for prediction intervals and are used for some metric computation.
Quantile 0.5 is always used (the median is required for some metrics). Quantiles can be defined with a precision up to four decimals.

Partitioned models

Partitioned model support for time series forecasting is experimental.