Interact | partitions | <partitionName> | offerserving | Built-in Learning Config
These configuration properties define the database write settings for built-in learning globally. This will be applicable for all partitions.
insertRawStatsIntervalInMinutes
- Description
The number of minutes the Unica Interact learning module waits before inserting more rows into the learning staging tables. You may need to modify this time based on the amount of data the learning module is processing in your environment.
- Default value
5
- Valid Values
A positive integer
aggregateStatsIntervalInMinutes
- Description
The number of minutes the Unica Interact learning module waits between aggregating data in the learning stats tables. You may need to modify this time based on the amount of data the learning module is processing in your environment.
- Default value
15
- Valid Values
An integer greater than zero.
autoAdjustPercentage
- Description
The value that determines the percentage of data the run of aggregation tries to process based on the metrics of the previous run. By default, this value is set to zero, which means the aggregator processes all staging records, and this auto adjustment functionality is disabled.
- Default value
0
- Valid Values
A number between 0 and 100.
saveOriginalValues
- Description
You can set the values as “All Values", “Binned Values", or “None". This will control what values will be logged in table UACI_LearningAttributeHist.
If “All Values" is selected then all learning attributes will be logged in the table. If this parameter is set to “Binned Values" then only those attributes will be logged in the table for which bins are created under “Interact-> Global Learning".
If set to “None" no values will be logged in UACI_LearningAttributeHist .
By default this is set to “None".
- Default value
None
- Valid Values
All Values | Binned Values | None
excludeAbnormalAttribute
- Description
The setting that determines whether to mark those attributes as invalid. If set to IncludeAttribute, abnormal attributes are included not marked as invalid. If set to ExcludeAttribute, abnormal attributes are excluded and marked as invalid.
- Default value
- IncludeAttribute
- Valid Values
- IncludeAttribute | ExcludeAttribute
smoothFactor
- Description
- When updating the recency weights of offers, which is a key factor for calculating the learning scores, the importance of the recent activities can be further optionally adjusted by comparing the number of recent activities to historic activities.
- Default Value
- 0.00
- Valid Values
- 0.00 to 1.00
- Example
- 0.01 | 0.001 | 0.05 | 0.1