Time Series Forecast - Data360_Analyze - 3.12

Data360 Analyze Server Help

Data360 Analyze
Product family
Product name
Data360 Analyze
Data360 Analyze Server Help
First publish date

Uses the Holt-Winters method to forecast a time series that can optionally contain a trend and seasonal variations.

Tip: Before working with this node, there are a number of prerequisite steps, see Working with the Statistical and Predictive Analytics nodes.
Note: An additional Statistical and Predictive Analytics node pack license is required to run this node. See Applying a node pack license.This node processes data in-memory. Additional RAM will be required when processing data sets with a large volume of data.

The node accepts the name of a field on the input pin that contains the univariate time series data to be modeled.

The node can be optionally configured with the type of time series data that is to be assumed when modeling the data. The time series characteristics can be set to indicate whether or not it contains a systematic trend and/or seasonal variations.

When the time series type is set to indicate the time series contains a seasonal component, the node can be configured to use an Additive method or Multiplicative method when decomposing the time series data.

The node can be configured with Frequency of the time series data, which indicates the number of sampling points that are included in the data for each time series period. For instance, the time series data may reflect data sampled with a quarterly, monthly, or hourly frequency. Arbitrary frequencies can be configured to indicate other intervals, e.g. a frequency of "100" could be used to indicate that the time series represents sensor data that has been sampled every one hundredth of a second.

The start period of the time series can be specified, e.g. for Yearly data the start could be "2010". Similarly, if the time series Frequency is configured to indicate there are multiple samples per period the start sub-period can also be set, e.g. for quarterly data, the start sub-period could be set to a value of 1,2,3 or 4.

The node can be configured with the number of future intervals (periods or sub-periods) to be forecast. At least one future interval must be forecast.

When run, the node uses the embedded R engine to "decompose" the time series data into its constituent parts and forecast the required number of intervals ahead.

The node's data output pin contains the decomposed time series components. The components output depend on the type of time series that was configured and includes values for:

  • The fitted value for the model.
  • The level component.
  • The trend component.
  • The seasonal component.



Specify the input field that contains the time series data to be forecast.

A value is required for this property. This property is case sensitive.


Optionally specify the characteristics of the time series data to be assumed when forecasting the time series. Choose from:

  • No Trend or Season - Data has no systematic trend or seasonal components.
  • Trend Non-seasonal - Data has a systematic trend but no seasonal components.
  • Trend and Seasonal - Data has both a systematic trend and seasonal components.

The default value is No Trend or Season.

If set to Trend and Seasonal the Frequency property must be set.


Optionally specify the seasonal method to use when performing a seasonal decomposition of the time series data using Holt-Winters filtering. Choose from:

  • Additive
  • Multiplicative

The default value is Additive.

This property must only be set when the SeriesType property is set to Trend and Seasonal and the Frequency property is set to a non-default value.


Optionally specify the frequency of the observations in the time series data in terms of the number of samples per unit of time. Choose from:

  • Yearly
  • Quarterly
  • Monthly
  • Hourly
  • Other Interval

The default value is Yearly.

If the SeriesType property is set to Trend and Seasonal, the default frequency Yearly cannot be used.


Optionally specify an arbitrary period to be used when the 'Frequency' is set to 'Other Interval'.

Must be an integer greater than 0.

If the SeriesType is set to Trend and Seasonal, the arbitrary period must be an integer greater than or equal to 2.

The default value is 1. This default value is equivalent to a Frequency of Yearly.


Optionally specify the start period of the time series data. Must be a numeric value e.g. '1996' for yearly data.

The default value is 1.


Optionally specify the sub-period of the time series data.

For example where the StartPeriod is a year, the sub-period may be the starting month for the data.

Must be a numeric value e.g. 11 for monthly data starting in November.

If specified, the StartPeriod property must be specified and the Frequency property set to Yearly, Quarterly, Monthly or Hourly.

If not specified, no sub-period is used.


Specify the number of periods in the future to predict. Must be an integer greater than 0.

A value is required for this property.

Inputs and outputs

Inputs: data.

Outputs: data.