Group locator weak keywords
Weak keywords are words that can have a different meaning based on the context. Strong keywords are words or phrases that always have the same meaning, independent of the context.
This feature is only required if you use this locator to create a Generic Knowledge Base, or if the training type is set to "Generic" or "Both", including documents directly trained for this locator with the generic algorithm. To train documents for the generic algorithm, set the training type to Generic or Both.
Select a dictionary that contains all weak keywords that are used during knowledge base creation and training. The dictionary needs one keyword per line, keywords in a phrase separated by a space, and keywords that are not case-sensitive. Typically, store weak keywords from different languages in the same dictionary file.
For each group locator, a predefined dictionary of weak keywords is installed. The dictionary contains keywords for US English, UK English, German, Spanish and French. For additional languages, update the dictionaries with keywords for those languages.
Locator | Dictionary file name |
---|---|
Invoice Group Locator |
IGL_WeakKeywords.txt |
Order Group Locator |
OGL_WeakKeywords.txt |
Amount Group Locator |
AGL_WeakKeywords.txt |
The concept of the generic algorithm is based on finding keywords that are in the same relative distance to a numeric value as displayed on a training document.