Documentation

holtWinters() function

holtWinters() applies the Holt-Winters forecasting method to input tables.

The Holt-Winters method predicts n seasonally-adjusted values for the specified column at the specified interval. For example, if interval is six minutes (6m) and n is 3, results include three predicted values six minutes apart.

Seasonality

seasonality delimits the length of a seasonal pattern according to interval. If the interval is two minutes (2m) and seasonality is 4, then the seasonal pattern occurs every eight minutes or every four data points. If your interval is two months (2mo) and seasonality is 4, then the seasonal pattern occurs every eight months or every four data points. If data doesn’t have a seasonal pattern, set seasonality to 0.

Space values at even time intervals

holtWinters() expects values to be spaced at even time intervales. To ensure values are spaced evenly in time, holtWinters() applies the following rules:

  • Data is grouped into time-based “buckets” determined by the interval.
  • If a bucket includes many values, the first value is used.
  • If a bucket includes no values, a missing value (null) is added for that bucket.

By default, holtWinters() uses the first value in each time bucket to run the Holt-Winters calculation. To specify other values to use in the calculation, use aggregateWindow to normalize irregular times and apply an aggregate or selector transformation.

Fitted model

holtWinters() applies the Nelder-Mead optimization to include “fitted” data points in results when withFit is set to true.

Null timestamps

holtWinters() discards rows with null timestamps before running the Holt-Winters calculation.

Null values

holtWinters() treats null values as missing data points and includes them in the Holt-Winters calculation.

Function type signature
(
    <-tables: stream[A],
    interval: duration,
    n: int,
    ?column: string,
    ?seasonality: int,
    ?timeColumn: string,
    ?withFit: bool,
    ?withMinSSE: bool,
) => stream[B] where A: Record, B: Record
For more information, see Function type signatures.

Parameters

n

(Required) Number of values to predict.

interval

(Required) Interval between two data points.

withFit

Return fitted data in results. Default is false.

column

Column to operate on. Default is _value.

timeColumn

Column containing time values to use in the calculating. Default is _time.

seasonality

Number of points in a season. Default is 0.

withMinSSE

Return minSSE data in results. Default is false.

minSSE is the minimum sum squared error found when optimizing the holt winters fit to the data. A smaller minSSE means a better fit. Examining the minSSE value can help understand when the algorithm is getting a good fit versus not.

tables

Input data. Default is piped-forward data (<-).

Examples

Use holtWinters to predict future values

import "sampledata"

sampledata.int()
    |> holtWinters(n: 6, interval: 10s)

View example input and output

Use holtWinters with seasonality to predict future values

import "sampledata"

sampledata.int()
    |> holtWinters(n: 4, interval: 10s, seasonality: 4)

View example input and output

Use the holtWinters fitted model to predict future values

import "sampledata"

sampledata.int()
    |> holtWinters(n: 3, interval: 10s, withFit: true)

View example input and output


Was this page helpful?

Thank you for your feedback!


Introducing InfluxDB Clustered

A highly available InfluxDB 3.0 cluster on your own infrastructure.

InfluxDB Clustered is a highly available InfluxDB 3.0 cluster built for high write and query workloads on your own infrastructure.

InfluxDB Clustered is currently in limited availability and is only available to a limited group of InfluxData customers. If interested in being part of the limited access group, please contact the InfluxData Sales team.

Learn more
Contact InfluxData Sales

The future of Flux

Flux is going into maintenance mode. You can continue using it as you currently are without any changes to your code.

Flux is going into maintenance mode and will not be supported in InfluxDB 3.0. This was a decision based on the broad demand for SQL and the continued growth and adoption of InfluxQL. We are continuing to support Flux for users in 1.x and 2.x so you can continue using it with no changes to your code. If you are interested in transitioning to InfluxDB 3.0 and want to future-proof your code, we suggest using InfluxQL.

For information about the future of Flux, see the following: