Configuring classifications

A classification is a grouping of documents that share common characteristics, such as word frequency.

Notes:

• Create at least two classifications before proceeding to training and production.
• For optimal performance, create no more than 100 classifications.

1. From the Embedded Web Server, access the configuration page for the application.

2. From the Classifications section, add or edit a classification.

3. Type a unique classification name.

   Note:  If necessary, specify the minimum confidence level. By default, the minimum confidence level is 90 percent.

4. Specify the number of pages for each document that you want to classify.

   • Single Page —Each document has only one page, and classification is done only on that page.

   • Constant Page Length —Each document contains the specified number of pages. Classification is done only on the first page, and the following pages are appended to the document.

   • Variable Page Length —The number of pages vary for each document. In a page length, the first page is classified and the succeeding pages are appended. For example, if 3 is the page length, then the first page is classified and the succeeding two pages are appended. If the pages are not recognized, then those pages are appended to the previously classified page until the application classifies another page successfully.

5. Configure at least one classification destination.

   Network destination

   1. Select Network Folder Destination .

   2. In the Network Folder Address field, type \\ myServer \ myFolder , where myServer is the server host name or IP address and myFolder is the folder name.

      Notes:

      • Make sure that the destination folder is shared and has a valid network address.
      • To set a dynamic folder name, in the address field, type % before and after the extraction field name. For example, \\ myServer \ myFolder \% fieldname % .

   3. Type the domain name and user credentials.

   E-mail destination

   1. Select E-mail Destination .

   2. Type the e-mail subject and e-mail address.

      Note:  To set a dynamic e-mail subject, in the address field, type % before and after the extraction field name. For example, \\ myServer \ myFolder \% fieldname % .

6. Select the output file format, and then type a name for the document.

   Note:  When you add an extraction field, the extraction field buttons appear following “Filename preview.” Click these buttons to add a dynamic file name. For multiple extraction fields, type an underscore between each name. For example, % fieldname1 %_% fieldname2 % .

7. If necessary, select Classification Summary E-mail .

   Note:  If the subject field is blank, then the classification name is used.

8. If necessary, select Audit XML Index File .

   Note:  By default, this option is selected. For audit purposes, the application creates an XML file for each scanned document in a classification. To disable this feature, clear the check box.

9. Add up to four extraction fields, and then specify the appropriate field names, types, and confidence threshold value. For more information, see Understanding extraction fields .

   Note:  In each extraction field, you can enable and apply autocorrect rules. To set up more rules, see Setting up Autocorrect rules .

10. Click Apply .