WikiTimes's Knowledge Extraction and Enrichment Process

The Wikipedia Current Events portal is a special part of Wikipedia that focuses on news events. Every day, a new page is added to this portal and news events of the day are collaboratively summarized and listed on the page. This is a very interesting part of Wikipedia that however has not been exploited so far by existing knowledge bases. In this technical paper , we detail the automatic knowledge extraction and enrichment process on which our knowledge base WikiTimes is built. WikiTimes information derived through this process is very high quality, with an accuracy of about 97-100%. In addition, we describe heuristics for enriching extracted events with complementary information that is not directly provided in the portal such as location and categorization. Using these heuristics, we are able to provide enriched information with an accuracy of about 93-96% for location extraction, 91-94% for categorization and up to 98% for related story.