Analyze and interpret results
The results of your experiment will appear in the Experiments tab.
This tab has two sub-tabs:
- Statistics to view your experiment statistics.
- Activity log to check logs.
The Statistics sub-tab shows the experiment results. On the top left you can see general information about the experiment, like:
- Experiment name.
- Experiment date and time.
- The number of analyzed documents.
- The number of documents with annotations and extractions.
- The library name.
- The ML model type.
- The match strategy, namely the criteria employed to match annotations and extractions.
Hover over the experiment name and select the pencil icon to edit the experiment name.
Watch the Results coverage panel to see the number of documents with extractions and the number of resources employed for the annotation so far with their respective percentages.
The Documents sub-panel lists in a table mode all documents with their respective number of class values.
It is possible to filter the list of documents by clicking Only documents with extractions according to your needs.
|All documents—with and without extractions—are visible|
|All documents with extractions are visible|
|All documents without extractions are visible|
Select the expanding arrow to see the classes that occur in the text.
- Select the N. of classes column header to sort the documents according to the number of class values.
The Resources sub-panel lists all the classes—grouped and ungrouped—of your project and for each class the total number of hits and documents where these classes occur (with percentage).
- Select Show card view to view your classes in a card view.
- Select Show table view to view your classes in a table view.
- When in card view, select Sort to sort your classes according to name and hits.
- When in table view, select the Name and Total hits column headers to sort your classes.