How News Evolves? Modeling News Text and Coverage using Graphs and Hawkes Process
This work addresses the challenge of combining natural language processing with dynamic system models for news monitoring, which is incremental as it applies existing graph and Hawkes process methods to a new domain with a custom dataset.
The authors tackled the problem of automatically monitoring news content over time by modeling news text as a sequence of directed multi-graphs representing semantic triples and fitting a discrete-time Hawkes process to capture dynamics. They demonstrated that this approach can infer whether articles were published by major or entertainment news outlets using a novel dataset.
Monitoring news content automatically is an important problem. The news content, unlike traditional text, has a temporal component. However, few works have explored the combination of natural language processing and dynamic system models. One reason is that it is challenging to mathematically model the nuances of natural language. In this paper, we discuss how we built a novel dataset of news articles collected over time. Then, we present a method of converting news text collected over time to a sequence of directed multi-graphs, which represent semantic triples (Subject -> Predicate} ->Object). We model the dynamics of specific topological changes in these graphs using a set of multivariate count series, which we fit the discrete-time Hawkes process. With our real-world data, we show that the multivariate time series contain both dynamic information of how many articles/words were published each day and semantic information of the content of the articles. This yields novel insights into how news events are covered. We show with the experiment that our approach can be used to infer from a sequence of news articles if the articles were published by major or entertainment news outlets.