Service Needs

Reference No.

N-0049

Title

Improve the auto-matching of two types of electronic documents originated from independent information source

Business Needs / Challenges

Two different types of electronic documents will undergo a matching process by a computer program to identify outstanding submission every day.  Contents of the common input fields in these two document types will be compared character by character for matching.  For those unmatched document records, officers have to refine the matching criteria, sometimes based on typo correction, to perform manual on-line matching again.  For records remained outstanding after going through the auto- and manual matching, officers will issue notice, either electronically or in paper format, to the party concerned for follow-up.  The project aims to enhance the existing system by using artificial intelligence, machine learning or other technology to identify the outstanding submission.

Application Areas

Commerce and Industry

Expected Outcomes

The enhancement should be able to minimize the human effort in the manual matching process.  With the enhanced system, the whole process can be shortened to boost efficiency and increase the matching rate.

Technologies to be Used

Artificial Intelligence (AI)
Deep Learning
Machine Learning
Natural Language Processing