Extraction online learning requires the data capture of all fields, including the correct field position on the document. Without field position and coordinates, the specific training algorithm cannot train the documents correctly. Use the mouse to capture the field position with the Point and Click option in Validation. To work with the keyboard, use auto completion to enter values and specify field coordinates at the same time.

Typing characters for the field value generates a small list that contains all possible words from the document starting characters you enter. The items from the list box are also highlighted on the document image. Typing more characters reduces the number of items in the list. The user can continue typing, the field is assigned the coordinates of the first value on the document that matches the field content.

Instead of typing, you can select an item from the list and confirm it by pressing Enter. This is useful for long invoice numbers or VAT IDs. If the same value is printed multiple times on the document, additional effort is required, for example, when Subtotal and Total are equal.

Auto completion, lists items according to the order in which they appear on the document page. The first matching recognition result appears first in the list. Use the arrow keys to select the correct value from the list on the document page in the Document Viewer and in the mini viewer. For extraction online learning, select the list item with the correct field position. The wrong field position can cause incorrect extraction results for other documents with the same layout. If you do not use extraction online learning, the exacted field position is irrelevant.