How can I find the frequency for the Russian word "чем-нибудь"?

Question

I've tried searching for the word "чем-нибудь"on the 20,000 word list I found on Wikipedia. You can link directly to it here. For the sake of thoroughness, I even searched the larger 50,000 word list that you can find here. And, I have also tried searching for it with the largest corpora available for Russian on Sketch Engine — the ruTenTen: Corpus of the Russian Web — to no avail. I also tried searching for this word on educalingo. Again, nothing.

On a more promising note, I did a Google search for this word and, apparently, this word shows up on 4,900,000 web pages. So, it isn't as if this is some rare word I'm searching for. I also ran the word through Google's Ngram and at least it was recognized there:

Is there any way to determine word frequency — something that assigns a rank to the word —

by using what I've found via the Google Ngram? I would think that there would be, but I'm not a mathematician, so I'd love to know how a linguist would go about doing this. Just an idea. I'd be just as happy to run this word through a list or database and have it give me its rank compared to all the other words in Russian.

Answer 1

Asking the Leipzig Wortschatz Projekt gives the answer:

Word: чем-нибудь Number of occurrences: 5,018 Rank: 33,962 Frequency class: 14

Information on the corpus:

Information on corpus: Russian Mixed 2013 Description

Russian mixed corpus based on material from 2013 Details Name rus_mixed_2013 Sentences 118,550,890 Language Russian () Types 14,388,948 Genre Mixed Tokens 1,800,364,710 Year 2013 Link to the corpus

https://corpora.uni-leipzig.de?corpusId=rus_mixed_2013

Answer 2

Using the Araneum Russicum III corpus (I am linking to the readily available 125Mtokens subcorpus; after a (free) registration you get access to the whole huge 19Gtokens one; the results are from the 19Gtoken corpus):

• что-нибудь (nominative): 25.20 ipm
• чем-нибудь (instrumental): 3.50 ipm
• ч.*-нибудь (should take care of all the cases, link to the frequency table by word form): 33.60 ipm

Note that the corpus is lemmatized, but the lemmatizer does not recognize such compound words; thus the shortcut with the regular expression. Also note that there are other (incorrect) variants in the corpus, e.g. "чем нибудь", "чем - нибудь" etc. And do not forget there is also "чём-нибудь" often written without the "ё".

Also note that raw frequency (or instances per million) can be misleading if one author (newspaper, web portal etc.) overuses the word. This should not happen with such a frequent expression, but very definitely can happen with rare words. ARF is available from the Word list (divide by 19778.053615 to get ipm):

• что-нибудь 275,770.70
• чего-нибудь 41,755
• чем-нибудь 40,644.70
• чему-нибудь 7,172.40
• чём-нибудь 1,875.20

Answer 3

The online version of Russian Nation Corpus (in Russian) offers frequency statistics of word forms, broken down by year:

https://processing.ruscorpora.ru/graphic.xml?env=alpha&mode=graphic_main&mycorp=&mysent=&mysize=&mysentsize=&mydocsize=&dpp=100&spp=&spd=1&text=lexform&sort=i_year_created&g=i_year_created&lang=ru&nodia=1&startyear=1800&endyear=2019&smoothing=3&req=%D1%87%D0%B5%D0%BC-%D0%BD%D0%B8%D0%B1%D1%83%D0%B4%D1%8C

You can also download the raw frequency stats from here:

https://ruscorpora.ru/new/ngrams/1grams-3.zip

and find out that the word form чем-нибудь ranks 4716th.