THE HLTCOE APPROACH TO THE TREC 2012 KBA TRACK
Brian Kjersten, Paul McNamee · 2012
Our team submitted runs for the TREC KBA Cumulative Citation Recommendation task. This task involves labeling over 300 mil-lion documents for whether they are relevant and/or central to particular entities already in a database. For this task, we used an SVM classifier that uses unigrams and named en-tities as binary features. In this paper, we de-scribe our work for the 2012 evaluation and the results we obtained. 1