Starting Automatic Classification in the Classification Inbox 
You can start automatic classification of documents to be classified in the classification inbox. This is only useful if you know that certain documents were not classified into suitable categories when the last automatic classification took place.
In an example-based taxonomy, this occurs if you create or train new categories, or if you change the training set for existing categories.
In a query-based taxonomy, this occurs if you create new categories and define new search queries, or if you change search queries for existing categories.
Note
You can use the procedure described here in the following scenarios, for example.
Example 1: You have assigned a new repository with documents on the topic Sport. You have created a new taxonomy for Sport. You have then manually classified some of the new sport documents in order to train the new categories. You now want to classify all remaining unclassified sport documents into the new categories automatically.
Example 2: You have already carried out an automatic classification run. However, some documents were not classified since insufficient training documents were available. You have therefore manually classified some of the unclassified documents in the classification inbox, thereby improving the training of the categories. You now want to classify all remaining unclassified documents automatically.
If you want to start an automatic classification run of all documents in the classification inbox, proceed as follows:
In the left-hand section of the screen, choose Documents to Classify.
Choose Auto Classify.
The system carries out the automatic classification in the current taxonomy of all documents that are in the classification inbox (Documents to Classify).
In the left-hand section of the screen, choose Documents to Classify.
Select a folder from the Path field.
Select documents from the list, and choose Auto Classify.
The system starts the automatic classification run as a separate process. The automatic classification process can take a few minutes to complete. You can monitor the current status in the crawler monitor. For more information, see Crawler Monitor.