Master Thesis

The problem

Together with a partner company in Norway, we are developing on a document management system as part of our project portal. The idea is that a user can upload multiple documents at once and that the system recognizes the name of these files and auto matches these with existing files in this folder. The user has to either agree with the choice the algorithm has made or make the ”right decision”.  There is no set naming convention for these documents as they are coming from multiple sources, and this would be very hard to enforce. So, we decided that the user has to match the documents manually. This, of course, can be tedious and can be improved.

The challenge

We are looking to make the algorithm smarter by learning from the choices users make when selecting a document, they want to match. Gather as much data as possible, find correlations and patterns in the choices that are made to propose better matches in the future. This will give our users a better user experience, will make our projects more efficient and will help us with the implementation of this system.

You will be working closely with our partner in Norway, Catenda. The CTO and developers will be your sparring partner and will guide you through the existing codes and assist you with setting up the machine learning tool. Communications can go through Teams and they are eager to let you help them in the development of this tool.