<< Click to Display Table of Contents >> Current:  Advanced Analytics > Algorithm  Holt-Winters

Used to analyze a numerical variable that changes over time, and predict future variables based on historical data.

 

How-To-Use:

The pre-node of this node needs to have the output of the data set, and there is at least one numeric field to be analyzed in the data set, and a date and time field Numerical fields may have two changes over time: one is trend, the other is periodic This model can be used to meet the above characteristics When the data has only trend characteristics, [Trend component] is set to add or mul, [Periodic component] is set to None When the data has only periodic characteristics, [Trend component] is set to None, [Periodic component] is set to add or mul When the data shows trend and periodicity at the same time, [Trend component] is set to add or mul, and [Periodic component] is set to add or mul.

After setting Holt-Winters, you can view the performance indicators by connecting to the table view; connecting to the picture view to view the time series analysis and prediction results.

 

Precautions:

Only one column can be selected for numeric column Time column can only select one column

Time column cannot have duplicate values

[Backward Prediction Span] and [Span Unit] are used in combination, for example: [Backward Prediction Span] is set to 10, [Span Unit] is set to M, which means 10 months of backward prediction [Span Unit] Set as the time interval between each row of data.

❖Configuration

After adding the Holt-Winters node to the experiment, you can set the Holt-Winters node through the "Configuration" page on the right.

[Trend component] add: additive effect; mul: multiplicative effect.

[Periodic component] add: additive effect; mul: multiplicative effect.

[Period] Period setting of data.

[Alpha] Smoothing index.

[Beta] Trend component coefficient.

[Gamma] Periodic component coefficient.

[Backward Prediction Span] Fill in an integer number n to indicate backward prediction n steps.

[Span Unit] Y: year; M: month; D: day; W: week; h: hour.

[Time column] A field whose data type is date and time.

[Numerical column] Numerical field that changes with time.

Right-click menu of Holt-Winters:

❖Run Holt-Winters Node

Run the node, pass the data to DM-Engine for calculation, and get the output result.

 

❖Reset Holt-Winters Node

The node that has been running is reset, the returned result is deleted, and the node status is changed to not running.

 

❖Rename Holt-Winters Node

In the right-click menu of the Holt-Winters node, select "Rename" to rename the node.

 

❖Refresh Holt-Winters Node

In the right-click menu of the Holt-Winters node, select "Refresh" to update the synchronization data or parameter information.

 

❖Save as Composite Node

In the right-click menu of the Holt-Winters node, select "Save as Composite Node",The selected node can be saved as a composite node to realize a multiplexing node, and the parameters of the saved node are consistent with the original node.

 

❖Cut Holt-Winters Node

In the right-click menu of the Holt-Winters node, select "Cut" to realize node cutting operation.

 

❖Copy Holt-Winters Node

In the right-click menu of the Holt-Winters node, select "Copy" to realize node  replication operation.

 

❖Delete Holt-Winters Node

In the right-click menu of a Holt-Winters node, select "Delete" or click the delete key on the keyboard to delete the node and its input and output connections.