Online Learning tab - Properties of Line Item Matching Locator window

The Line Item Matching Locator engine can enhance the extraction results by unassisted training, requiring no setup or knowledge sources beyond standard configuration options.

General

This group has the following options:

Automatically generate training data

Select this option to run online learning during batch processing.

This option is cleared by default.

Number of documents required to repeat incremental learning

Set the threshold number of documents collected before training begins. The files used for incremental training are kept until they are used by the permanent training step. This option is only available when the "Automatically generate training data" option is selected. The value for this option is set to 100 by default.

Number of documents required to repeat permanent learning

Set the threshold number of training documents collected before the information is transformed into a permanent model file. The higher the number of documents specified, the higher the accuracy as well as memory consumption. All training documents that are saved for permanent learning are then deleted from the collection path. This option is only available when the "Automatically generate training data" option is selected. The value for this option is set to 1000by default.

Maximum number of training documents to be gathered

Set the threshold number of documents trained before the system ceases training. This option is only available when the "Automatically generate training data" option is selected. The value for this option is set to 10000 by default.

Use training data

Select this option to use the trained data during extraction. This option is cleared by default.

Definitions for the buttons at the bottom of this window can be found in Common Project Builder Buttons.