CLAN analysis question (combining freq and combo?)

[Alice Wang]

Hi,

Is it possible to use the freq command with strings containing multiple words?

For context, I had been using the combo command to search up the words that came before and after certain token words and was wondering if there was a way to tally up the frequencies of unique frames surrounding those tokens.

I would greatly appreciate any suggestions!! 

Thanks,

Alice W. 

[Brian MacWhinney]

Dear Alice,

    I'm trying to think of an example that might illustrate what you are searching for, but I am coming up with too many possible patterns.  Could you please give me the exact combo command you are using and the files or perhaps one sample file on which you are running it.  Also, a brief sketch of what the tallied output might look like, based on your command and the sample file.

-- Brian MacWhinney

--
You received this message because you are subscribed to the Google Groups "chibolts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/6c7a9f89-c966-4542-91fe-b7bfe6325da4o%40googlegroups.com.

[Leonid Spektor]

Alice,

Yes it is possible to track adjecent words. For example, try command(s):

For speaker tier:

freq +s"not nice"

For %mor tier either one of:

freq +sm;not^;nice

freq +sm"|neg |adj"

freq +sm|neg^|adj

freq +sm|neg^;nice

Leonid.

[Alice Wang]

Hi Brian,

Yes – I had been submitting the command:

combo +t*MOT +s"_*^ball^_*" +u *.cha

I've also attached a sample input file (and the output I get from the combo command). I was hoping to get an output with the frequencies of the different frames surrounding the word "ball" (e.g. the ball here (100), blue ball is (50), etc.) Please let me know if you have any suggestions, and thank you for your help!! 

Best,

Alice 

On Friday, June 5, 2020 at 11:54:06 AM UTC-7, macw wrote:

Dear Alice,

    I'm trying to think of an example that might illustrate what you are searching for, but I am coming up with too many possible patterns.  Could you please give me the exact combo command you are using and the files or perhaps one sample file on which you are running it.  Also, a brief sketch of what the tallied output might look like, based on your command and the sample file.

-- Brian MacWhinney

On Jun 5, 2020, at 1:31 PM, Alice Wang <afuz...@gmail.com> wrote:

Hi,

Is it possible to use the freq command with strings containing multiple words?

For context, I had been using the combo command to search up the words that came before and after certain token words and was wondering if there was a way to tally up the frequencies of unique frames surrounding those tokens.

I would greatly appreciate any suggestions!! 

Thanks,

Alice W. 

--
You received this message because you are subscribed to the Google Groups "chibolts" group.

To unsubscribe from this group and stop receiving emails from it, send an email to chib...@googlegroups.com.

[Alice Wang]

Hi Leonid, 

Thank you for this! I was wondering if there was a method for tracking multiple words when not all of the tokens are predefined (for example, if I wanted to find the frequency of unique "frames" (words preceding and following) for the word "nice." Sorry for not clarifying that earlier, and I would appreciate any input!

Thanks,

Alice 

On Friday, June 5, 2020 at 11:55:42 AM UTC-7, Leonid Spektor wrote:

Alice,

Yes it is possible to track adjecent words. For example, try command(s):

For speaker tier:

freq +s"not nice"

For %mor tier either one of:

freq +sm;not^;nice

freq +sm"|neg |adj"

freq +sm|neg^|adj

freq +sm|neg^;nice

Leonid.

On Jun 5, 2020, at 13:31, Alice Wang <afuz...@gmail.com> wrote:

Hi,

Is it possible to use the freq command with strings containing multiple words?

For context, I had been using the combo command to search up the words that came before and after certain token words and was wondering if there was a way to tally up the frequencies of unique frames surrounding those tokens.

I would greatly appreciate any suggestions!! 

Thanks,

Alice W. 

--
You received this message because you are subscribed to the Google Groups "chibolts" group.

To unsubscribe from this group and stop receiving emails from it, send an email to chib...@googlegroups.com.