Create custom dictionary from character vector

Question

I am trying to look for specific words in corpus using dfm_lookup().

I am really struggling with the dictionaries needed for the dfm_loopup().

I created a character vector named "words" which contains all the words that should go into the dictionary.

dictionary needs a list, so I am creating a list from the character vector before I am using dictionary().

dict <- dictionary(list(words))

But then I get

Error in validate_dictionary(object) : 


 Dictionary elements must be named: digital digital-tv digitalis ...

What do I have to change in the list command to get the proper output for dictionary()?

Is there a simplier version to look for specific words in a dfm? Because it was really easy with the tm() package.

As documentation and error suggest you need to name elements of the lists: **dictionary(list(words = words))** That should help. If not try reading carefully https://www.rdocumentation.org/packages/quanteda/versions/1.4.1/topics/dfm_lookup — Krzysztof Nowicki, Mar 18 '19 at 09:43

score 2 · Accepted Answer · answered Mar 18 '19 at 09:45

2

I believe you need to name the items in a list in order to use dictionary with quanteda. Here is an example:

library(quanteda)

words = c("cat","dog","bird")

word.list = as.list(words)
names(word.list) = words

dictionary(word.list)
Dictionary object with 3 key entries.
- [cat]:
  - cat
- [dog]:
  - dog
- [bird]:
  - bird

answered Mar 18 '19 at 09:45

NM_

1,887
3
12
27

Thank you very much, that helped me a lot! – BanffBoss122 Mar 18 '19 at 10:19
@BanffBoss122: You're welcome, glad I was able to help :) – NM_ Mar 18 '19 at 15:15
The input `x` in `dictionary(x)` needs to be a named list, as stated in `?dictionary`. So `dictionary(list(mykey = c("value1", "value2")))` etc. – Ken Benoit Mar 18 '19 at 22:00

Create custom dictionary from character vector

1 Answers1