Search for Articles:

Recently Downloaded: Paper ID: ART20171512, Total 176 Articles Downloaded Today

Improved Information Extraction with NLP & CRF for Invoices

Monalin Pal

Abstract: In a general Invoice processing system, the process starts from manually uploading the different format of invoices like paper, pdf, excel, document, image etc. for handling a supplier invoice, from its receipt to upload in the ERP system and is ready for payment. This invoice process is manually handled for any organization. Each process of manual upload and payments was time consuming. This is an attempt to use AI in invoice processing system, where we can process invoices automatically by extracting the relevant information from the invoices and feeding it to the ERP systems through RPA (Robotic Process Automation). This process will help to reduce costs while improving both the accuracy and the speed of data extraction. Invoice Processing Advisor that can read invoice and extract the accurate information in a representable format. Annotations and Machine Learning for data extraction from the preprocessed data. Intelligent Invoice Processing will help to cater the increasing business needs as we see that majority of organizations spend lot of man FTE?s and time for processing invoices. With the help of AI, this solution will be helpful to extract the information from the invoices more accurately and reduce the manual effort as well as the cost. We can easily scale it to multiple type of invoices as well introduce various business rules check that can communicate with the RPA?s to perform any type of complex actions.

Keywords: Artificial Intelligence, CRF Model, Natural Language Processing, Python

Country: India, Subject Area: Computer Engineering

Pages: 1237 - 1239

Edition: Volume 8 Issue 3, March 2019

How to Cite this Article?

Monalin Pal, "Improved Information Extraction with NLP & CRF for Invoices", International Journal of Science and Research (IJSR),, Volume 8 Issue 3, March 2019, 1237 - 1239

Download PDF

Viewed 62 times.

Downloaded 22 times.