.../data/salience/confidence

This directory contains files that are examples of data files for implementing entity and entity type confidence functionality. Users can customize the following files

confidence_entity.dat

A set of queries that can be used to increase confidence in the extraction of specific entities, see below

confidence_type.dat

A set of queries that can be used to increase confidence in the entity type assigned to a specific entity, see below

These files may be customized within a confidence section of a user/salience directory, using the files from the default directory as a model.

confidence_entity.dat

This contains the confidence queries for entities (based on label). An entity can be mentioned several times in confidence_entity.dat if you wish but it will have to pass every confidence query specified for it to be marked as confident. The columns for confidence_entity.dat are:

entity<tab>confidence query

You can add an optional 3rd parameter to an entry in confidence_entity.dat that enables you to change the label assigned to the entity based on the results of the query.

George Bush<tab>silicon AND "chip design" AND engineer<tab>Not the President

would get an entity of 'Not the President', if you got a confidence match.

confidence_type.dat

This contains the confidence queries for entity types. A type can be mentioned several times in confidence_type.dat if you wish but it will have to pass every confidence query specified for it to be marked as confident. The columns for confidence_entity.dat are:

entity type<tab>confidence query