Skip to main content

CES

This document describes the usage of complex exponential smoothing (CES) for forecasting.

Description​

CES uses complex-valued state-space equations to model level, trend, and seasonal changes in a time series. TDgpt uses automated CES modeling to select an appropriate model for the input time series and forecast future data.

CES provides confidence intervals for forecast results.

Parameters​

ParameterDescriptionRequired?
periodThe number of data points included in each period. If not specified or set to 0, a non-seasonal model is used.No
modelCES model type. Enter N, S, P, F, or Z. The default value is Z, which automatically selects a model.No

N specifies a non-seasonal model, S specifies simple seasonality, P specifies partial seasonality, F specifies full seasonality, and Z automatically selects a model.

Input data must contain only finite numeric values. A non-seasonal forecast requires at least two input values. When period is greater than 0, the input data must contain at least two complete periods.

Example​

In this example, forecasting is performed on the i32 column. Each 12 data points form a period, and CES automatically selects the model.

FORECAST(i32, "algo=ces,period=12,model=Z")

The complete SQL statement is shown as follows:

SELECT _frowts, FORECAST(i32, "algo=ces,period=12,model=Z") from foo
{
"rows": rows,       // Rows returned
"period": period,   // Period of results
"algo": "ces",     // Algorithm
"mse": mse,         // Mean square error (MSE)
"res": res          // Forecast results and confidence intervals in column format
}

References​