Information density, Heaps’ Law, and perception of factiness in news
Miriam Boon
ACL Workshop on Language Technology and Computational Social Science (ACL LACSS 2014)
Baltimore, Maryland, USA, June 26 - 26, 2014
Abstract
Seeking information online can be an exercise in time wasted wading through repetitive, verbose text with little actual content. Some documents are more densely populated with factoids than others. The densest documents are the most efficient use of time, likely to include the most information. This study explores this problem using crowdsourced ratings of the factual content of 772 online articles. The results suggest that after controlling for widely varying document length using Heaps' Law, a significant positive correlation persists between perceived factual content and relative information entropy.
START
Conference Manager (V2.61.0 - Rev. 3312)