Mining Events from Wikipedia

Last semester I presented progress on mining texts for descriptions of events by looking for statistically significant co-occurrences of dates and names. This semester I will present progress on mining descriptions of events from a rather more structured source: Wikipedia chronologies. Wikipedia has a great many chronology or timeline articles that are rich sources of 1 or 2 sentence event descriptions. By scraping these articles and parsing the individual chronology entries into event representations, using the Wikipedia links as a high-quality form of named entity detection, I can quickly assemble databases of events. I have been experimenting with making these events available on the web as Linked Data and queryable via SPARQL.