You are here

MA Florian Kunneman

Automatic Detection and Prediction of news events.

My PhD is aimed at the prediction of the news. That is to say, nobody knows exactly what the front pages of tomorrow will be like, but there are certain types of news events that tend to have an overall pattern that can add to expectations. For example, in the case of a strike, reports about or demonstrations of social discontent might have been displayed earlier. Likewise, celebrity divorces are often anticipated upon in gossip media.

Large amounts of news messages, extracted from sources ranging from tweets to newswire, will be linked to specific event types. The resulting event types and their accompanying news articles will be used in order to train a classifier, that will be able to predict whether an event will take place given a new input of news articles.

The automatic prediction of news is useful for news agencies, as it gives them a tool to react quicker to the course of a news event. The task of the PhD is situated in text mining and machine learning, and is closely related to event detection and trend prediction.

Works with