Intelligent document processing based on RPA and machine learning
Xufeng Ling, Ming Gao, Dong Wang
Abstract
RPA (robot process automation) is an intelligent tool that used to automatically process the business documents. Partially, incorporating human input into the learning process helps to improve the efficiency over 200% while the error could be controlled within 0.5%. To this effect, firstly, we applied OCR to the document (either PDF or picture) to identify the specific element; secondly, the text features of the document are extracted by natural language processing technology; and at the end, an automatic description and classification are proposed using machine learning algorithm. We also incorporated the human-computer interaction into the proposed system for modification and validation of the extracted features. Indeed, the proposed model transforms the experience of document administrators into artificial intelligence algorithm through RPA document processing, which greatly reduces the influence of leave and job transfer of the document administrators, and solves the about problem of document processing and meets the needs of the work.