Configure Classification Jobs
You can create classification jobs to automate detection of languages and classification of documents when the documents are received. After classification jobs are committed, asset languages are set in Oracle Content Management. The results can then be used to create appropriate custom digital assets and assign the correct language to the source asset.
Note:
If your Oracle Content Management instance was created before mid-February 2023, you need to enable OCI services content sharing deployment options required for using advanced Content Capture features. For more information, see Edit Your Oracle Content Management Instance in Administering Oracle Content Management.- In the procedures pane on the left, select your procedure.
The configuration pages for the selected procedure appear on the right.
- Open the Processing tab.
-
In the Classification Jobs table, click
, or to edit an existing job, click
.
You can also copy a classification job by selecting it, clicking
, and entering a new name when prompted. Copying a job allows you to quickly duplicate and modify it.
-
On the Document Selection page:
- In the Job Details section:
-
Enter a name and description for the job.
- Select the Online check box to make the available for processing.
- Select Language Detection and Document Classification actions. Actions is a required field. One of the two actions must be selected to proceed further. However, for efficiency, it is recommended that you create a single job to automate language detection and document classification.
-
- In the Document Processing section:
- Select the Process Documents check box if applicable to the choices you made in the previous steps.
-
To process the classification job for specific document profiles, select one or more document profiles listed in the Restrict to Document Profiles field, or select All to process documents for all defined document profiles. Not selecting any is also equivalent to All.
- If you selected the Document Classification action check box
in the Job Details section, the Attachment Processing section is
enabled for you:
- Select the Process Attachments check box if applicable to the choices you made in the previous steps. This option leverages auto language detection and auto classification for attachments that are received from Content Capture Client.
- Restrict to the required attachment types by selecting the check boxes for the available attachment types. You can also select all of them.
- In the Job Details section:
- If you selected the Language Detection action in the Job Details
section on the Document Selection page, the Language Detection page is
enabled for you:
- Selecting primary language is mandatory. Select the required language from the Primary Language drop-down list. Primary language has the maximum weightage in the document. This is also the language that is the most prominently used in the document.
- Set the minimum threshold by dragging the Minimum Threshold bar. It isn't not mandatory to set the threshold as it will at least be zero. The minimum threshold ties with the primary language. The threshold indicates that if there is no language above this value, the Primary Language field remains blank.
- Select the required languages from the Document Language drop-down list. A document can have multiple languages.
- If you selected the Document Classification action in the Job
Details section on the Document Selection page, the Document
Classification page is enabled for you. On this page:
- In the Document Classification section:
- Define classifications types out-of-the-box by document understanding in the Document Profile Mappings section, if required.
- Set the minimum threshold to indicate the confidence
score for classification types. This score applies across all
classification types. The minimum threshold applies to the
mappings. This threshold must be met and the mapped document
profile is assigned.
However, this also pertains to the classification type that has the highest value. For example, document understanding may give the type of invoice with the highest value. If that is mapped and above the threshold, then the document profile is assigned. If there are no matches or the threshold isn't met, then the default document profile is used.
- In the Attachment Classification section,
- Define classifications types out-of-the-box by document understanding in the Attachment Type Mappings section, if required.
- Set the minimum threshold to indicate the confidence score for classification types. This score applies across all classification types. The minimum threshold applies to the mappings. This threshold must be met and the mapped attachment type is assigned. However, this also pertains to the classification type that has the highest value. If there are no matches or the threshold isn't met, then the default attachment type is used.
- In the Document Classification section:
-
On the Post-Processing page, specify based on the following what happens after your classification job completes:
- No system error situations are cases in which all the criteria on the previous page were met. A successful transformation can flow to commit.
- System errors are any cases in which the transformation fails: no records found, too many records found, and so on. For unsuccessful transformations, the batch returns to the Content Capture Client for repair.
-
Review settings on the Summary page and click Submit to save the job.
-
Configure how batches flow to your classification job. See Configure Batch Flow to a Classification Job.
-
Test the classification job you created.
Configure Post-Processing and Monitoring of a Classification Job
Use post-processing options of a classification job to specify what happens after processing completes.
Configure Batch Flow to a Classification Job
Deactivate or Delete a Classification Job
When you delete a classification job, it no longer remains available for batches for which it is set as a post-processing step. If a job specified for post-processing is not available, an error results for the batch. You may want to change a job to offline for a time before deleting it, allowing you to resolve unexpected issues with its deletion. Online classification jobs run when they are selected in a client profile or on the Post-Processing page of a processor job. You can temporarily stop a job (take it offline) or change a deactivated job to run again. You cannot delete batch processing jobs if they are configured as a post processing job in another batch processor.