Within the framework of the European research project CPN (Content Personalisation Network), we are now launching the third and final pilot study on personalised news.
Nowadays, the supply of news is so limitless that making choices has become a necessary evil. With VRT MYNWS, the name speaks for itself, we offer VRT NWS readers customised news. Last year we launched a similar call to gather testers for this innovation project. Initially we targeted 500 testers. In the end, just over a thousand test candidates enrolled. This year we hope to convince even more VRT NWS app users to participate in this test pilot.
Hybrid recommendation system
With VRT MYNWS you get a personal newsfeed based entirely on your interests, in addition to the top stories and the most recent articles from the VRT NWS editors.
The test application uses three algorithms to create personalised content.
First of all, the app registers which articles, image or audio clips you click and uses artificial intelligence to select similar messages for you. This is called content-based filtering. Do you click on domestic news and political opinion pieces over and over again? Then you will receive suggestions in the same vein. Your preferences can of course fluctuate. Maybe the Corona virus has sparked your non-existent interest in virology. As your reading behaviour changes, the software of VRT MYNWS will adjust the suggested content.
A second way to predict what you’ll probably like to read is collaborative filtering. The principle is pretty simple: an algorithm divides people into groups based on what they read, listen to and watch. You are assigned a digital profile that shows certain similarities with others. The category of like-minded souls in which you end up determines what is presented to you. People with the same news consumption can unconsciously influence what you see on your screen. If you and person X have seven news items in common, chances are that the last two articles that person X read will also interest you.
Does this sound technical? The combination of the above methods is actually similar to how Netflix and Spotify make recommendations. So most of you are already using such filtering systems on a daily basis.
An often mentioned point of criticism about news personalisation is that readers (unintentionally) end up in a bubble. VRT MYNWS takes this into account and integrates a third algorithm offering you random articles. In this way you keep on being surprised and you are occasionally taken out of your reading, viewing and listening comfort zone. In this manner, we try to pierce the filter bubble.
Your own news
What is the added value of an app that brings 'your' news? We suspect that convenience will be one of the most obvious answers. A second advantage is that you can avoid articles which you normally have to scroll past. But we prefer to hear that feedback from you. The more tips that contribute to the further development of this software, the better.
Would you like to join this experiment? Sign up via the VRT Pilootzone.
Together with Italian, French, British, German, Greek and Cypriot partners from the tech and media industry, VRT Innovation has developed software that gives news organisations the opportunity to inform their audiences in a more targeted way. At the end of this project we intend to launch this technology on the market.
Would you like to know more about the CPN project? Go to https://www.projectcpn.eu.