The importance of negative filters for document digitisation
In Doc² there are two options for document classification. You can classify documents using AI. Here we use a neural network as well as TF-IDF in Doc². The negative filters are important to get the best results.
You probably use our AI for 90% of the document types. But there is also the case where you define the document type via a regex. In the regex, you define that this document type is taken if a certain combination of characters or words occurs in the document.
In a few exceptions, however, you will also need a negative filter that allows you to exclude certain things.
You have the possibility to open the negative filters in the Document Validation view:
There you can define the filters and apply them via Apply:
Documentation: Exception filters