Within the PROVIDENCE and PROVIDENCE+ projects, IDLab is developing technology that enables news content providers to monitor and predict the reach of their offered online news, and to adapt their production and publication strategies in order to optimize the overall reach and impact.
The overall goal of these projects is to answer questions such as:
- Which types of stories work best on which platform and what time of day?
- What is the ideal strategy to disseminate news online, depending on the channel and/or platform used?
- How big will be the impact of a given article two hours after publication?
To accomplish these goals, different core competences of IDLab are combined:
- Big Data: a stable and easy extendable monitoring framework is set up and managed with IDLab technology on our own server infrastructure, monitoring and processing every read and related social activity in real-time.
- Machine Learning: combining machine learning techniques with mathematical models, accurate prediction engines are designed, predicting on every moment what the expected reach is of every news article on the different publication channels.
- Information Extraction: in order to design such accurate prediction models, knowledge about an article’s content is essential input for these models.
Within IDLab, many SOTA algorithms and techniques are developed for retrieving structured information out of unstructured data. Active research topics range from categorisation, keyword extraction or sentiment analysis to multi-axis clustering, similatity calculation or semantic enrichment.
Combining big data and machine learning techniques to find the optimal media strategy.