Statistical functions

FORECAST.ETS

Not implemented

Listed by Microsoft; no catalogue row yet. How we measure this

No local names are recorded for this entry.

What it does

FORECAST.ETS(target_date, values, timeline, [seasonality], [data_completion], [aggregation])

FORECAST.ETS predicts a future value from a timed series, using the AAA version of exponential smoothing.

FORECAST.ETS(target_date, values, timeline, [seasonality], [data_completion], [aggregation])

  • target_date is the point to predict, a date or number after the end of the history.
  • values is the history.
  • timeline is the matching dates or numbers, with a constant nonzero step. It need not be sorted.
  • seasonality is 1 to detect the pattern (default), 0 for none, or a positive pattern length up to 8,760.
  • data_completion is 1 to fill missing points with the average of their neighbours (default), or 0 to treat them as zeros. Up to 30% of points may be missing.
  • aggregation combines values sharing a timestamp: 0 AVERAGE (default), then SUM, COUNT, COUNTA, MIN, MAX and MEDIAN.

Try it

Input cells (optional)

Example: {"B2": 3, "B3": 4}. A1 holds the formula. Other cells need values before a formula can read them.

Enter a formula, then calculate.

Runs locally in your browser with OneCalc. Its answer may differ from Excel. Use English function names and commas between arguments.

The OneCalc preview has no answer for this function: see the state above.

Quirks

  • #NUM!: the target precedes the end of the history, no constant step exists, or seasonality is invalid or above 8,760.
  • #N/A: values and timeline differ in size.
  • #VALUE!: Microsoft's timeline section lists this for duplicate timeline values, while its aggregation section says duplicates are combined. The two statements conflict, and no reading is recorded here to settle it.

How to implement

No implementation exists yet. It needs the AAA exponential smoothing fit, the detection of a constant timeline step, seasonality detection, the data-completion and aggregation rules, and recorded Excel answers to fix the optimiser's stopping rules. Forecast values depend on those rules, so they must be measured and not assumed.

Rust code

This entry has no Rust implementation yet.

No implementation has been published with a Handbook test suite for this function.

Version differences

Microsoft says these functions are not available in Excel for the web, iOS or Android. No Excel answers for this function are recorded here, so no comparison between Excel builds exists.

Alternatives

  • FORECAST.LINEAR for a straight-line forecast.
  • TREND and GROWTH for fitted series.
  • The Forecast Sheet command, which builds on the same family.

Named variants

No named variants are recorded yet.

Suggest a correction or contribute an implementation · For agents