Configure Classifier Window

Configures a Classifier Definition

How to Launch

Can be launched only from Create Classifier Window.

User Interface

Toolbar

• Browse Allows user to open a Classifier Definition File.

• Reset Discards the changes in current definition from last save.

• Save Changes Saves the changes in current definition file

Files

Allows the user to select sample files for the configuration. The selected sample file will be the Current Document and it will be displayed in the document viewer.

• Browse Button Allows user to select the sample folder for each Class.

• OCR All Button Applies OCR on all sample documents

Document Viewer

Displays the Current Document. If the current document is a PDF document or Image document , it will be displayed in Image Document Viewer. If the document has OCR applied, then the OCR text will be displayed in a Text Document Viewer in another tab. If the document is a text document, it will be displayed in Text Document Viewer.

Class Configurator

Allows the user to configure the classifier definition.

Toolbar

• Apply Button Classifies the Current Document using the definition.

• Apply All Button Classifies all the documents using the definition.

Classification results are shown in the Classification Results Pane.

The configuration UI depends on the type of classifier used.

Keyword Classifier

• Add Button Adds a new Keyword.

• Included Toggle Button Sets whether to Include/Exclude the Keyword in matching.

• Delete Button Deletes the Keyword

Two types of Keywords can be added in Keyword Classifier.

1. Regex Uses Regular Expressions for finding the matching keyword. It uses PCRE as the regex engine. Uses can apply following Regex options :

   • IgnoreCase

   • SingleLine

   • MultiLine

   • Unicode

   • Global

   • Sticky

2. Word Uses words for finding the matching keyword. More than one word can be provided separated by comma. Following options are available for modifying the matching :

   • Match Any Only one one of the words are needed to match.

   • Match Case Case sensitive search while matching.

For more information about Regular Expressions and to learn, reference, and test visit the following websites :

• Regex101

• RegExr

Classification Results

Shows the classification results for the Current Document or for all the documents.

• Document Name of the document

• Actual Class The class in which the document is added as a sample.

• Predicted Class The class as predicted by the classifier. This can be empty.

How to configure the Classifier

Keyword Classifier

Keyword classifier works by searching each keyword on a Class in the document. A class is considered a match if :

• All Included keywords are found

• All Excluded keywords are not found

see Keyword Classifier for more detailed overview

Classes are tested for match in the order they appear in the definition. Once a class is matched with a document other classes are ignored and the matched class is set as the result of the classification for the document. If none of the classes match with a document, its class will be null.