Modify Prediction Input Dialog

The Modify Prediction Input Dialog allows you to modify the settings associated with one or more prediction inputs.

Ä    Note: Optimizable functions do not use this dialog. Instead, they use a modified version of the Modify Field Dialog.

Ä    Note: Fields that are currently being used by the Solution Service cannot be modified.

&  For help with predictions, see Predicting and Modeling Financial Data.

Modifying Prediction Inputs

The following values are displayed on this dialog.

p    Field Name

This is the name of the field used for this input.

p    Data Series / Group

This is the data series or group which the field belongs to. In most cases, this will be the same data series or group in which the prediction is being created.

¨    Preprocessing

This value specifies the type of preprocessing that should be used on the selected field. Preprocessing normally defaults based on the type of value contained in the input field.

·      Percent Change

This option uses the percent change in the values from the previous bar as the input. It is typically used for continuous, positive, non-zero values that do not have a constant range, such as price information.

·      Change

This option uses the change in the values from the previous bar as the input. It is typically used for continuous values that may not have a constant range, such as some momentum indices.

·      None

This option uses the actual value as the input. It is typically used for discrete values that have a constant range, such as entry/exit signals.

·      Conditional

This option causes this value not to be used as an input to the prediction. Instead, it is used to determine which bars are included for training. It the value is non-zero, the bar is included. If the value is zero, it is skipped.

Ä    Note: If multiple conditional inputs are specified, all of the values must be non-zero for the bar to be included for training.

·      (Multiple Settings)

This option indicates that multiple inputs are selected and that they currently use more than one preprocessing setting. Leaving this value will result in these values remaining unchanged. Changing this value will result in all selected inputs being changed to the newly selected value.

Ä    Note: Change and percent change preprocessing are done relative to the previous bar of the prediction data series. In most cases, this is the change from the previous bar of the input. However, when using data of other periodicities, it will be the change from the value at the previous bar of the prediction.

&  For additional help with working with multiple periodicities, see Understanding Periodicity.

þ    Lag Input

This option allows you to lag the selected input. By default, the value of the selected input is taken for the current day. Selecting this option and specifying a lag greater than zero allows you to include values from previous days. When this option is enabled, you can add the multiple inputs as Discrete Inputs or as an Input Group. Discrete Inputs allow you to select, deselect, and optimize each input individually. Input Groups allows you to view, select, deselect, and optimize the inputs as a single unit.

:    Example: To use today’s value, yesterday’s value, and the previous day’s value of a field from a data series with daily data, select the Lag Input option and set it from 0 to 2.

Ä    Note: It is typically not necessary to lag inputs. By default, TradingSolutions uses Time-Lag Recurrent Networks which automatically maintain a memory of recent input values.

Ä    Note: Input lags are done relative to the prediction, not relative to the inputs. For example, for a prediction that updates every minute, an input with a 10-minute periodicity only changes once every ten bars. Therefore, most of the values in the lag will be either the current value or the last value. Lagging of these longer periodicity inputs should be done using calculations directly in the longer periodicity data series.

Ä    Note: Advanced neural network users may notice that using multiple samples for the inputs to a Basic Multilayer Perceptron has a similar effect to using a Time-Lag Recurrent Network with Focused TDNN memory. The primary difference between the two methods is that the Lag Input setting allows you to specify different memory depths, whereas focused TDNN memory only allows for only one memory depth to be used for all of the inputs.

Ä    Note: If multiple inputs are selected and they currently have different values for the minimum or maximum lag, these values will not be available for modification.

p    Split into separate inputs…

Inputs which are currently lagged across multiple values are optimized and/or enabled as one unit. To treat each lag independently, press the Split into separate inputs… button.

Ä    Note: This button is not available when multiple inputs are selected.

þ    Include in training.

This option enables or disables the input as being a part of the prediction. It duplicates the functionality of the individual field checkboxes on the Predict a Value Wizard: Select Inputs page and the Modify Field Dialog: Prediction Inputs page.

Ä    Note: Genetic optimization is the primary cause for an input being disabled.

Ä    Note: Predictions created in previous versions of TradingSolutions included disabled inputs to determine which inputs values were available for each given date. Support for this behavior is controlled from the Modify Training Settings Dialog. If disabled inputs are included, removing disabled inputs may affect your results.

Ä    Note: Conditional inputs cannot be disabled or optimized.

Ä    Note: If multiple inputs are selected and they currently have different values for this feature, the checkbox will be dimmed. Leaving the checkbox dimmed will result in these values remaining unchanged. Changing this value will result in all selected inputs being changed to the new setting.

þ    Allow optimization to include/exclude this input.

This option indicates whether an input can be enabled or disabled as part of optimization of the prediction. It allows genetic optimization to try turning on and off the Include in training checkbox.

Ä    Note: If multiple inputs are selected and they currently have different values for this feature, the checkbox will be dimmed. Leaving the checkbox dimmed will result in these values remaining unchanged. Changing this value will result in all selected inputs being changed to the new setting.

What Do I Do Next?

When you are done modifying the selected prediction input, press the OK button. If you would prefer to exit this dialog without making modifications, press the Cancel button.

How Did I Get Here?

The Modify Prediction Input Dialog appears when you press the Modify Input… button on the Predict a Value Wizard: Select Inputs page and the Modify Field Dialog: Prediction Inputs page