MOZAIQ’s Document Classification (Indexing) Solution leverages machine learning to become smarter the more it processes
Mortgage document classification is available out of the box, with over 280 documents already pre-trained using machine learning. The platform uses page level classification and rules for over 500 pages in a mortgage loan document, with accuracy levels above 90% once the platform has been trained using “real world” documents – the higher the number of indexed document type, the higher the accuracy, further enhanced to over 99% with our human in the loop offering.
Key features include:
- Enhanced OCR and Computer Vision Pipeline
- Entity detection – able to classify which borrower is listed in a document
- Can be trained with per-customer documents including cover sheets, etc.
- Configurable SLA based on expected document volume and size