Is there a minimum amount of words required on a document for intelligent keyword classification?
I’ve trained it on a specific document that has about 10ish words, but every time I use the Intelligent Keyword Classifier for this document, even after training over it, it returns nothing.
The Intelligent Keyword Classifier in UiPath is designed to identify document types by analyzing recurring words and patterns within documents. While there isn’t a strictly defined minimum word count for effective classification, documents with very few words—such as those containing only about 10 words—may not provide sufficient data for the classifier to accurately learn and recognize patterns. This lack of content can lead to the classifier returning no results, even after training.
For optimal performance, it’s advisable to train the classifier with documents that have a richer textual content, allowing it to detect and learn the distinctive features of each document type. If your documents are inherently brief, consider supplementing the training set with additional documents of the same type that contain more text. This approach can help the classifier build a more robust understanding and improve its classification accuracy.
For more detailed guidance on configuring and managing the Intelligent Keyword Classifier, you can refer to UiPath’s official documentation.
Yea I have it trained over 300-400~ documents, many are detailed but some are brief. 95% of the time it does classify documents, but sometimes it just doesn’t. The only thing I can think of is that limited word amount.