Mansi NLP Grammar

Finite state and Constraint Grammar based analysers, proofing tools and other resources

View the project on GitHub giellalt/lang-mns

mns meeting

Agenda

Status

fst

Fst went up from 0.95267 to 0.96176.

Suggestions

             typos      Avrg pos        % missp        % missp
             .txt       for corr        in 1st         in top-5     
-----------------------------------------------------------------
240410:       473        1.08          54.59          57.72        
240411:       473        1.08          53.42          56.41        
240422:       547        1.09          46.75          49.54        

Trond to discuss technicalities with Sjur (who is back home).

yamls

Let’s not forget them. The are for future meetings.

Missing words

Csilla has been working on Luima_Seripos_2013-2017.missing.freq.240208.

We now have 337 “TBS” (to be solved) in this file. In most of the cases Csilla is at loss

Ща̄нюмн яныг паща ла̄тыӈ ке̄тэ̄гум, вос таквитэ ӯргалытэ, 
mother+N+Sg+Lat+PxSg1 big greeting+N+Sg+Nom send+VxSg1, may herself.Acc guard+V+Act+Ind+Prs+ScSg3+OcSg
йигрищагум-ягпыгагум вос ёмщакв янмалты.
sister+N+Du+Nom+PxSg1-brother+N+Du+Nom+PxSg1 may well grow+V+Act+Ind+Prs+ScSg3

Jack to look at TBS and eventually send as much as possible back.

Speller

Csilla continues collecting typos, Trond looks at testing.

More texts

Trond has added the remaining LS to the corpus.

corpus-mns-x-closed/converted/news/luima_seripos/luima_seripos_no_23_2023.pdf.xml
corpus-mns-orig-x-closed/news/luima_seripos/luima_seripos_no_23_2023.pdf
corpus-mns-x-closed/converted/news/NAME.pdf.xml
corpus-mns-orig-x-closed/news/luima_seripos/NAME.pdf

We try to get a meeting with Børre this week. Csilla to systematize conversion errors and report for the meeting.

Next meeting

29th at 1400 Finnish time.