![]() | Define the Domain Attributes (Object Phrases Tab) | Review Accumulated Phrases | ![]() |
A sufficiently representative sample of messages should be used in identifying the Attribute Phrases to gain a good insight into the how their subject writers express themselves in conveying information.
This process is supported from within the Phrase Matching Tab of the TextProcessor Object editor. Using ODBC data sources, multiple data sets can be defined on the same data source by extracting different results based on a 'Where' clause. The 'Where' clause takes the same syntax as an SQL 'Where' clause, for instance
db_field1 = 'db_field1_value2' AND db_field2 <> 'db_field2_value3'
Once the data set has been selected the messages can then be read in from that data set for analysis. Individual messages can be selected in the Messages Window and their content reviewed in the lower Message Analysis Window to detect text expressions than can be mapped to Attributes already defined on the Object Phrases Tab. Phrases can be highlighted in this lower window and added to the appropriate Attribute using the Add button and the nearby drop down list box.