Enable Additional Context Checks

You can turn on additional context checks for some grammars, which can eliminate some false positive matches. This option adds a small performance penalty to the extraction speed, so it is turned off by default.

You can use this option for name grammars, and for grammars with numeric matches, such as ID, driving, TIN, passport, and health.

Name Grammar Context Checks

For the name grammar, for example, context checks mean that Named Entity Recognition rejects matches that are part of a larger string that is not a name. For example, it does not match M Records as a name when it matches in the string A&M Records.

To enable this feature, you can use the following configuration.

[Eduction]

ContextCharsBeforeMatch=5

ContextCharsAfterMatch=5

PostProcessingDefaultParameters=LuaParams



[LuaParams]

IDOL_PXI_NAME_CHECK_CONTEXT=TRUE

Alternatively, you can set IDOL_PXI_NAME_CHECK_CONTEXT as an environment variable.

Numeric Grammar Context Checks

For numeric grammars, context checks mean that Named Entity Recognition rejects matches that appear to be part of a longer token, such as a decimal number, hyphenated number, file name, or URL.

To enable this feature, you can use the following configuration.

[Eduction]
ContextCharsBeforeMatch=5
ContextCharsAfterMatch=5
PostProcessingDefaultParameters=LuaParams

[LuaParams]
IDOL_PXI_CONTEXT_CHECK=TRUE

Alternatively, you can set IDOL_PXI_CONTEXT_CHECK as an environment variable.