Summary Details

Query:	WebCorpus Counts
Author:	Jerry Kurjian
Submitter Email:	click here to access email
Linguistic LingField(s):	Text/Corpus Linguistics
Summary:	Regarding query: http://www.linguistlist.org/issues/16/16-1291.html#1 Below I summarize the comments of Andrew Kehoe and Antoinette Renouf (5/27/2005), two of the creators of WebCorp, who kindly replied to my query concerning WebCorp in thread 16.1291 and on Corpora list (corpora AT uib.no): Within a webpage, WebCorp will gather as many kwics per page as there exist, if the ''one hit per page'' option is not checked. Across webpages, WebCorp only gathers hits from up to 200 webpages. Getting fewer than 200 hits might mean that you have chosen to filter some out features out, that some of the 200 webpages were not accessible to WebCorp or had change, or that there are fewer than 200 pages that have the search term. Finally, the authors say they are continuing to upgrade WebCorp, and in an upcoming version plan to add frequency counts, type/token ratios, collocation profiles, and ''other statistics.''

LL Issue:	16.1366
Date Posted:	29-Apr-2005
Original Query:	Read original query
Back

Sums main page

New from Cambridge University Press!