Smart recommender technology for news providers

Tailor-made news

NewsTAPAS is an innovative system developed by ML2Grow, in collaboration with Roularta Media Group, VRT & Imec. This system will be able to adapt content to the needs and context of the user. A content recommender system is nothing new, but what makes NewsTAPAS unique is that it not only will allow news providers to automatically adapt the news messages to the profile of the user, but also to the type of device, time of day and location.

We want to increase the attention span of the user by tailoring their news experience and keeping an overload of information to a minimum. This will greatly increase your user experience.

NewsTAPAS is an innovative system developed by ML2Grow, in collaboration with Roularta Media Group, VRT & Imec. The NewsTAPAS project is a response to the growing demand for automation and personalisation of the way in which news items are displayed.

The problem

People feel overloaded with information and overwhelmed by digital distractions which reduces their attention span. This makes it harder to reach and retain an audience of readers. Recommendation systems currently approach this problem by selecting the news items that best match a user’s profile, but this does not hold the user’s attention.

Our solution

Our intention was to improve the user experience by helping increase the user’s attention span and keeping information overload to a minimum.

Each publisher stores articles in a standardised way in its content repository. The different parts of each article are classified (e.g. the title, introduction, by-line, paragraphs, photo) and can be used to show the article in different ways to users. This is based on information such as the type of device used, time of day, interests and reading the history of the user. We turned to machine learning to create a profile of readers based on their behaviour. We then determined which parts of an article best meet their needs. Our recommender system allows you to predict which articles will be most relevant to a reader depending on how much interest a reader shows in a particular type of new article. The system can adapt content to the needs and context of each user. While this type of system is not new, NewsTAPAS is unique because it enables news providers to not only adapt their news messages to reader profiles, but also to the type of device used and the time of day and location.

Work carried out

Recommender system – identifies the most relevant information in a text and rewrites the content in a more concise form

How we added value

News content is automatically adapted – according to the type of device, time of day and location of the reader

Raised the attention span of users

Enhanced the user experience