Extracted Input Features (OUTFEATURE)

_seasonalDummyi

Indicator for season i

indicates whether the time period belongs to the season denoted by i. The i indicator ranges from 1 to the highest level for seasonality. The seasonality range depends on the time interval set for the time variable. See Seasonality for more information.

• 0 — The time period does not belong to the season denoted by i.
• 1 — The time period belongs to the season denoted by i.

This variable is not included if Seasonal dummy variables is unchecked for stage 2 feature extraction.

_TREND_

Predicted values from the trend model

predicted value from the trend model. This variable is not generated if Dependent variable trend is set to None for the neural network model.

_Predict_ESM

Predicted values from the best exponential smoothing model

predicted value from the best ESM model. This variable is not generated if you uncheck ESM forecast of dependent variable for the neural network model.

lagYi

dependent_variable_name lag i

value of the dependent variable lagged by i time periods. The i indicator ranges from 1 to the number of lags that you specify for the dependent variable for the neural network model in stage 2.

_lagXiP_independentVarName

independentVarName lag i

value of an independent variable lagged by i time periods. The i indicator ranges from 1 to the number of lags that you specify for Number of lags for the independent variables for stage 2 feature extraction. If this value is 0, then this variable is not included.

_roleVar

Data partition indicator

value indicating the data partition to which this observation belongs. The data partition is specified with the following values.

• 1 — fitting (in-sample) partition
• 2 — validation (holdout) partition.

  Observations with value 2 appear only if you specify a nonzero value for the holdout options.

• 3 — test (out-of-sample) partition.

  Observations with value 3 appear only if you specify a nonzero value for the Number of periods to exclude from modeling option in project's forecast settings.

Observations in the forecast horizon have a value of 0 to indicate that they play no role in model fitting, validation, or testing.

_nnetTargetVar

Target variable for stage 2

target variable for the neural network model in stage 2

_seasonalDummyi

Indicator for season i

indicates whether the time period belongs to the season denoted by i. The i indicator ranges from 1 to the highest level for seasonality. The seasonality range depends on the time interval set for the time variable. See Seasonality for more information.

• 0 — The time period does not belong to the season denoted by i.
• 1 — The time period belongs to the season denoted by i.

This variable is not included if Seasonal dummy variables is unchecked for stage 2 feature extraction.

_lagXiP_independentVarName

independentVarName lag i

value of an independent variable lagged by i time periods. The i indicator ranges from 1 to the number of lags that you specify for Number of lags for the independent variables for stage 2 feature extraction. If this value is 0, then this variable is not included.

_roleVar

Data partition indicator

value indicating the data partition to which this observation belongs. The data partition is specified with the following values.

• 1 — fitting (in-sample) partition
• 2 — validation (holdout) partition.

  Observations with value 2 appear only if you specify a nonzero value for the holdout options.

• 3 — test (out-of-sample) partition.

  Observations with value 3 appear only if you specify a nonzero value for the Number of periods to exclude from modeling option in project's forecast settings.

Observations in the forecast horizon have a value of 0 to indicate that they play no role in model fitting, validation, or testing.

_regTargetVar

Target variable for stage 2

target variable for the regression model in stage 2