- Full text PDF available (2)
Data Set Used
This paper describes an approach submitted to the 2013 PAN com-petiton for the source retrieval sub-task. Three different methods for extracting queries were used, which employed tf-idf, noun phrases and named entities, in order to submit very different queries and maximize recall.
This paper describes an approach submitted to the 2014 PAN competition for the source retrieval sub-task . Both independent term and phrasal queries are generated, using either term frequency-inverse document frequency or noun phrases to select the terms.