View and Edit Features

View and edit the linguistic features and categories of features (feature sets) that Tiberias should take into account when analyzing the text classes in your experiment.

Deciding which feature sets to select and possibly which features within that feature set to remove depends on the goals of your experiment.

This page enables you to:

• View data about each of the features in the text classes you defined, for each of the currently selected feature sets.
• Edit your selection of feature sets and of individual features.
• Recalculate the expected accuracy of the classifier.

To open the View and Edit Features page:

In your experiment, on the Analysis of Classes page, click View/Edit Features.

The top 250 features of each class are displayed.

Element	Description
Active Categories	The currently selected feature sets.
Feature	Click a feature name to view all examples of the feature within the text classes you defined. The full scriptural verse in which the feature appears is displayed for each occurrence of the feature, with the feature highlighted. Note: The Examples page displays the examples grouped according to the text class or test text in which they were found. Click a navigation link at the top of the Examples page to see the associated examples. Click the info icon next to a lexeme or a disambiguated word to view a translation and its part of speech.
Category	The feature set to which the feature belongs.
Class	The text class to which the feature belongs.
Frequency	The frequency of occurrence of the feature within the text classes (%). To see the actual number of occurrences, hover over the percentage number.
Occurrences in all Classes	The number of occurrences of the feature in all text classes.
Significance (Based on T-Test)	The significance of the feature in determining classification.
Update	Saves all of the changes you have made and recalculates the expected accuracy.
Search	Search the features in your text classes. Note: Because the View and Edit Features page displays the top 250 features of each class, it is possible that a search term that appears in one of your text classes will not be located because it is not significant enough.