The keywords list in AntConc is, as the name suggests, a tool to create a list of keywords. To do this your target corpus is compared to a reference corpus. The target and reference corpora do not need to be of the same size. The comparison is then done statistically. The statistics in AntConc used for this task are either chi-squared and log-likelihood.
In AntConc load your corpus or corpora. Go to Wordlist tab then click start.
Select the Tools Preference menu.
Click the ‘Keywords List’ option, then click ‘Add Files’.
Check the desired file is there. Click ‘Load’ then click ‘Apply’.
Go to ‘Keyword List’ tab then click ‘Start’.
A list of types should appear like this.
The keywords are ranked by default by the keyness. In this example the top ranking type in “english” with a score keyness (in this example, chi-squared) of 729.913 (this is a combined score of both the target and reference type score). And it has a frequency of 822 in the target list.
You are more than welcome.
LikeLiked by 1 person
Do I need to choose first a Corpus, say Brown, and then go to Kewords list and tool preferences and load a different corpus, say LOB?
LikeLike
Yes. The list can be a full list (another entire corpus) or a list with the frequencies created (a text file with types and their frequencies).
LikeLike
Hey,
I would like to know if it is possible you sent me the wordlist of the keywords from Portuguese national corpus. I am searching about it and I can not download the whole corpus.
LikeLike
I am sorry but I only work with English and Japanese corpora. I don’t know of any Portuguese corpora or anybody working with them.
LikeLike
Hi,
thank you for your useful explanation. But please help me understanding the “keyness”. What is it exactly? How is the keyness score calculated? You say, it is a combined score of both the target and reference type score. I don’t really understand that.
LikeLiked by 1 person
There are different ways to calculate keyness. Two of the most common are chi-squared
https://en.m.wikipedia.org/wiki/Chi-squared_test
and log-likelihood
https://reference.wolfram.com/language/ref/LogLikelihood.html
Essentially they both take the frequencies of the target words, size of the corpora and compare these across the corpora.
LikeLike
Reblogged this on TESOL_Peter and commented:
A nice explanation of AntConc and its Keyword list function by Warren Tang.
I started using AntConc again for some qualitative research and discovered this while searching for advice. Thanks Warren for posting!
LikeLiked by 1 person
Peter,
Thanks. Glad this post was of use to you.
LikeLiked by 1 person