A Machine Learning System for Tracking Sentiment in Irish Economic News

URL:

Abstract:

Tracking sentiment in the popular media has long been of interest to media analysts and pundits. With the availability of news content via online syndicated feeds, it is now possible to automate some aspects of this process. There is also great potential to “crowdsource” much of the annotation work that is required for the construction of predictive models.

Here we introduce a system that uses active machine learning techniques to mon- itor sentiment in economic news from popu- lar Irish media sources. The proposed system (outlined in Figure 1) has two novel aspects. Firstly, it generates an aggregated news feed containing a diverse set of articles relevant to the Irish economy. This allows users to browse articles in their preferred news reader application, and annotate these articles as “positive”, “negative” or “irrelevant” via embedded links. Secondly, the results of this manual annotation process are used to train a supervised learner that labels a much larger set of news articles. The annotation and classification trends can subsequently be tracked online.

• Tagged
• XML
• BibTex