Re: [antconc:2118] Problem with keyword list

[Laurence Anthony]

Hi Wonhee,

I think you might be confused by the way the keywords work. Load in a reference corpus in the keywords tool preferences (raw text or a list) and then click Start in the keywords tool. AntConc will calculate the frequencies of your target corpus, calculate the keywords, and show the results.

I hope that helps.

Regards,

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

On Wed, 7 Nov 2018 at 09:14, WON HEE Yee <wonh...@gmail.com> wrote:

Dear Mr. Anthony,

I was making a keyword list with ANC Written Corpus. But the following two problems were generated.

    1) The result of wordlist coincided with keyword list in terms of frequency and the ranking. (Please see the second attached image of AntConc word list).

    2) Keyness values in the keyword list were in four digits, i.e. +1994.58, +1994.58, 1462.04, ... (Please see the first attached image of AntConc keyword).

        Usual values are 'considered' to be in two digits, i.e. 91, 85, ..., only if I'm not mistaken. :-)

The above results have been made by AntConc 357. But those problems led me to try the lower versions of AntConc, i.e. 354, 324, ... But all the attempts were met with the same problems.

Could you possibly inform me what should be done to solve the problems? I'm in need of help. Thank you.

Best

Wonhee Yee

Seoul

South Korea

--
You received this message because you are subscribed to the Google Groups "AntConc-Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antconc+u...@googlegroups.com.
To post to this group, send email to ant...@googlegroups.com.
Visit this group at https://groups.google.com/group/antconc.
For more options, visit https://groups.google.com/d/optout.

[WON HEE Yee]

Dear Mr. Anthony

Thank you very much for your time for the reply.

Now I've done the work the way you said, and the results are shown in the attached two images. I found the two problems described in my below message were still existent -  the same list as word list and extraordinarily high values of keyness)

I followed the process: loading the reference corpus --> Click 'Start' in the keyword tool

Of course, ANC Corpus for reference was loaded. Also, the stop list was loaded in order to eliminate the function words such as articles, modal verbs, be verbs.

With these problems, I'm in a bit of a trouble submitting the keyword list.

Can you help?

Best regards

Won Hee Yee

[WON HEE Yee]

Dear Mr. Anthony,

Thank you for the reply.

By the way, one problem still remains, which is too high keyness values (please see the keyness values in the attached image). How do I solve the problem?

Could you help? Thank you again.

Best regards

Won Hee Yee

p.s. I need to clear the problem in order to complete my Ph.D. dissertation.

On Wednesday, November 7, 2018 at 9:20:35 AM UTC+9, Laurence Anthony wrote:

[Laurence Anthony]

I am 100% certain that your reference list is not being loaded correctly. Make sure that the format matches the format produced by the word list export option. Also, make sure you have not made a mistake with the selection of raw texts vs a word list.

If you continue to have problems, send me a screenshot of the keyword tools preferences window.

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

[WON HEE Yee]

Dear Mr. Anthony,

Thank you very much for the reply. I did as you advised, and found I had a mistake of selecting a word list instead of raw text. Thank you indeed.

The attached screenshot is the keyword list with keyness values much lower than the previous one.

But I found the keyness is still high - three digits; previously they were in four digits. 

My question is whether these values are acceptable in terms of corpus analysis? I think I've seen values in two digits in many cases.

I would appreciate it if you answer the question. Thank you.

Regards

Won Hee Yee

p.s. In the 2nd and 3rd attachment, I'm also sending the screenshot of tool preferences of wordlist and keyword. 

[WON HEE Yee]

Dear Mr. Anthony,

Problems here are similar to the initial one.

  1) Results of wordlist and keyword are identical as shown on 1st and 2nd attached screenshots. But this time, raw text in wordlist tool preferences is selected as in the 3rd attachment. (Screenshot of keyword tool preference is also in 4th attachment.)

  2) In addition, keyness values of the keyword list are still high - in three digits, as shown in the 2nd attachment.

Could you review the two problems and tell me what should be done about the problem. Thank you very much.

Regards,

Won Hee Yee

On Wednesday, November 7, 2018 at 11:22:27 PM UTC+9, Laurence Anthony wrote:

[WON HEE Yee]

Mr. Anthony,

I almost forgot ... the attached keyword list also shows that the ranking of frequency and keyness drops from the highest to lowest. I am suspicious about the result because the keyword has been produced in sorting by keyness, not by frequency. In that case, the ranking of frequency keeps changing between drop and rise. Have I misunderstood the result? Please help. Thank you.

Won Hee yee

[Laurence Anthony]

Hi,

The keyness values look fine to me. You often see the highest ranked keywords with values like these.

But, the fact that your ANC list is called "wordlist" and you're using the raw files setting seems very odd to me. Are you sure that you are using the correct files? Raw files are just plain text files with normal text in them (this paragraph is an example of *raw text*). Word lists are formatted like the output of the AntConc word list tool.

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

[WON HEE Yee]

Hi Anthony,

Thank you again.

In the attachment, I'm sending two sets of keyword list - ANC list in the setting of raw files and setting of word list.

Differences between the two settings are:

  1) Keyword with ANC list in the setting of raw files setting: keyness values are ok, but the frequency and keyness are in parallel sequence.

  2) Keyword with ANC list in the setting of word list setting: keyness values are high again (in 4 digits), but keyword frequencies are in the sequence of drop and rise, looking more natural.

According to your explanation, is it appropriate that I report the keyword list in the second attachment?

Thank you.

Won Hee

[Laurence Anthony]

Can you just send me your ANC list?

Lsurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

[WON HEE Yee]

Certainly.

Please find ANC list in the attachment.

By the way, thank you so much again for all your time and effort.

Best

Won Hee

[JFlorian]

Hi Wonhee Lee,

Every message you send is coming through twice, one as part of the larger thread, and another single post in a new thread.   If a post continues a topic, it's best to keep them together.

Also, when replying, it helps everyone if you trim or eliminate all prior responses.  An easy way to make sure old replies are not resent in a new post is to hit Reply or Reply to All, put your cursor in the message box, highlight all (control key plus A) and hit delete.  Type your message as usual and send.

Judy

[WON HEE Yee]

Dear Mr. Anthony,

I would rather streamline the questions regarding keyword list.

The first and second attachments of keyword list look fine in terms of ranking and sequence. But keyness values are still high - in three and four digits.

Regarding this high keyness values, you asked me to send you the reference corpus, which is ANC written word list in this case.

So I am sending the three attached files.

I am so sorry to keep you busy with the questions.

Could you look at the files and keep me updated with possible solutions?

Thank you very much.

Best,

Won Hee Yee

Seoul

South Korea

[Laurence Anthony]

Hi,

The reference word list looks fine, and the keyword list values look fine, too. I don't see any problem with these results. The values are not particularly high.

Regards,

Laurence.

###############################################################
Laurence ANTHONY, Ph.D.
Professor of Applied Linguistics
Faculty of Science and Engineering
Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
E-mail: antho...@gmail.com
WWW: http://www.laurenceanthony.net/
###############################################################

--

[WON HEE Yee]

Dear Mr. Anthony,

Thank you very much for the reply.

Your review is critically helpful to my research. I feel grateful for your time on this issue.

Will be in touch for further issues in the future.

Thank you again.

Regards,

Won Hee Yee

+82 10 9155 8037

Skype: wonhee173