We are extracting account numbers from invoices. Some account numbers have spaces in them, some do not. In the data labeling session it appears as though the whole number is being picked up by the label but when the ML Skill is deployed and the data extracted, we are only receiving the first few numbers before the first space in the account number.
Example of Acct # (all parts are highlighted/captured for labeling in the data labeling session under field name ‘account-number’):
123 456 789
What we receive after ML Skill deployed and data extracted:
123
Only the first three numbers are returned.
Ideally the account number would come out as 123456789 (with no spaces).
Has anyone else run into this issue and found a solution?
In the data labeling session, make sure that the entire account number, including spaces, is consistently labeled. If the labeling is inconsistent, it may impact the model’s ability to recognize the complete account number.
We grouped all numbers in the line using the ‘/’ key and marked the field as multi line. This appears to be working to pull the full account/meter number.
