Adaptive Model form
Completing the Settings tab

Field

Description

This setting corresponds to the value that specifies the amount of interaction results history, which are translated in number of cases, the scoring models maintain during predictions. By default, it is set to never discard information (0).

The memory configuration allows you to discard the oldest cases, and it allows you to implement trend detection by creating multiple adaptive models, all triggered by the same proposition but with different memory settings. This setting influences the binning of predictors as behavior changes with new cases being recorded.

• Low memory values allow the identification of new trends.
• High memory values provide robust and long-term predictive power.

Field

Description

The value that determines the number of interaction results that trigger running data analysis for a model. Data analysis is triggered after the number of interaction results configured in this setting is reached.This setting should be configured according to the resources available to the ADM system and taking into account the minimum set of responses required for models to evolve. Default setting is 500.

The grouping settings work in conjunction to control how predictor grouping is established. The fact that a predictor has more groups typically increases the performance, but the model may become less robust.

• Grouping Granularity: a value between 0 and 1 that determines the granularity of predictor groups; higher values result in more groups, lower values in less groups. This setting establishes the threshold for a statistical measure that indicates the distinctive behavior between predictors groups. If the measure is above, the groups indicate significant distinctive behavior, otherwise they will be collapsed. Default setting is 0.25.
• Grouping Minimum Cases: a value between 0 and 1 that determines the minimum percentage of cases per interval. Higher values result in decreasing the number of groups, which may be used to increase the robustness of the model. Lower values result in increasing the number of groups, which may be used to increase the performance of the model. Default setting is 0.05.

A value between 0 and1 that determines the threshold for excluding poorly performing predictors. Default setting is 0.52.

A value between 0 and 1 that determines the threshold for excluding correlated predictors. Default setting is 0.8.

Field

Description

A value that determines the number of cases of moving window size per proposition. The number of cases of moving window size per proposition influences the calculation of the CoC, and it is implemented so that equal comparison between models can be performed. Default setting is 0.

A value that determines the number of interaction results that trigger refreshing the scoring models in the ADM database. Model refresh is performed when the number of interaction results in this settings is reached.

You should set this value to a value lower than the value for running data analysis. Default setting is 150.

Check to enable or disable updating the model's local profile after every response. This setting allows you to enable local (PRPC) learning for the adaptive models configured by the adaptive model rule. The default setting is enabled.

Check if you want adaptive model details captured in the work object's history. Default setting is disabled.