|
 |
Albanian with natural language processing | ||
Share with:
| ||
 Language Learning Forum : Language Learning Log |  | |
| Eginhard Pentaglot Newbie France enno-hermann.blogspo Joined 4492 days ago 12 posts - 23 votes Speaks: German*, French, English, Croatian, Lithuanian Studies: Albanian | Message 1 of 10 27 October 2014 at 5:23pm | IP Logged |
|
If you know Albanian, please don't correct me or give any hints, thanks!
Motivated by my university courses in natural language processing (NLP), I've recently started a new project to learn Albanian. I chose Albanian deliberately mainly because I know (almost) nothing about the language, but it's still distantly related to the other Indo-European languages, so what I have in mind should not be completely impossible. I also like the Balkans in general, having been there several times over the last 1.5 years and also learning Croatian. I'm planning to learn Albanian by using (almost) only Wikipedia. Specifically, I'll only use target-language resources and not consult any dictionaries or grammars. I've compiled a corpus of about 10 million words in about 0.5 million sentences from the Albanian Wikipedia. I will not simply read all articles, but instead use NLP techniques to find out e.g. the most common words, collocations of specific words in order to decipher vocabulary and grammar. This is mainly an experiment; I know very well that this is not a very efficient method to learn a language, I just want to find out what is possible this way. I have a blog (http://enno- hermann.blogspot.com/) where I'll be posting more general discoveries and more technical details (I'm implementing the NLP algorithms myself for practice) and will use this log to share smaller discoveries I make and to allow better discussion of my experiment. There is also a Google Spreadsheet where I compile my findings. Edited by Eginhard on 19 November 2014 at 12:28pm 9 persons have voted this message useful | |
| clmns01 Diglot Newbie Austria Joined 3493 days ago 22 posts - 23 votes Speaks: German*, English Studies: Portuguese, Italian | Message 2 of 10 27 October 2014 at 5:28pm | IP Logged |
|
Wow, this sounds like a crazy adventure! Viel Glück dabei!
1 person has voted this message useful | |
| Eginhard Pentaglot Newbie France enno-hermann.blogspo Joined 4492 days ago 12 posts - 23 votes Speaks: German*, French, English, Croatian, Lithuanian Studies: Albanian | Message 3 of 10 28 October 2014 at 2:27pm | IP Logged |
|
The abundance of geography stubs on Wikipedia already allowed me to identify some
basic sentence structures, several nouns and three cases which I'll call Nominative, Genitive and Locative. More details are also in the spreadsheet linked above. E.g. there are a lot of sentences like these: X is a city in region Y in country Z. X is the capital of Y. X has a surface of Y and Z inhabitants. ... Cases: Nominative: qyteti (city) Genitive: e qytetit (of the city) Locative: në qytetin (in the city) (But there are several declensions it seems) Edited by Eginhard on 28 October 2014 at 2:27pm 1 person has voted this message useful | |
 Iversen Super Polyglot Moderator Denmark berejst.dk Joined 6507 days ago 9078 posts - 16473 votes  Speaks: Danish*, French, English, German, Italian, Spanish, Portuguese, Dutch, Swedish, Esperanto, Romanian, Catalan Studies: Afrikaans, Greek, Norwegian, Russian, Serbian, Icelandic, Latin, Irish, Lowland Scots, Indonesian, Polish, Croatian Personal Language Map | Message 4 of 10 28 October 2014 at 5:18pm | IP Logged |
|
Wow, a field linguist among us! Good luck with your experiment.
1 person has voted this message useful | |
| Eginhard Pentaglot Newbie France enno-hermann.blogspo Joined 4492 days ago 12 posts - 23 votes Speaks: German*, French, English, Croatian, Lithuanian Studies: Albanian | Message 5 of 10 29 October 2014 at 10:25am | IP Logged |
|
Well, I have no training in field linguistics, although I find it very interesting
and this is a very nice intellectual exercise. But it's actually rather the opposite of traditional field linguistics I guess because I'm not in the field at all and already have a huge amount of data readily available. As regards Albanian, I learned the names of 7 months - those which are similar to the English names, the current one and February because it was easy to identify the one which never occurs with 30 or 31. I've grouped the remaining ones into 30 and 31-day months and will try to come up with some way to identify them other than looking up events of which I know the date. This would be a lot easier with a literary corpus where it should be possible to at least link months to their seasons. 1 person has voted this message useful | |
| Iversen Super Polyglot Moderator Denmark berejst.dk Joined 6507 days ago 9078 posts - 16473 votes Speaks: Danish*, French, English, German, Italian, Spanish, Portuguese, Dutch, Swedish, Esperanto, Romanian, Catalan Studies: Afrikaans, Greek, Norwegian, Russian, Serbian, Icelandic, Latin, Irish, Lowland Scots, Indonesian, Polish, Croatian Personal Language Map | Message 6 of 10 29 October 2014 at 3:09pm | IP Logged |
|
Your project is field linguistics insofar that you try to extract words and grammar directly from utterances in the language. And there is a lot of linguistic research into the best methods to do that - though mainly from the pre-Chomsky period (eg. the socalled 'Bloomfield school').
It is different from traditional field grammar insofar that you can't pose questions to the natives to clear up murky points. 2 persons have voted this message useful | |
| tristano Tetraglot Senior Member Netherlands Joined 3851 days ago 905 posts - 1262 votes Speaks: Italian*, Spanish, French, English Studies: Dutch | Message 7 of 10 31 October 2014 at 7:55am | IP Logged |
|
wooow, looking forward to see your progresses!
1 person has voted this message useful | |
| Eginhard Pentaglot Newbie France enno-hermann.blogspo Joined 4492 days ago 12 posts - 23 votes Speaks: German*, French, English, Croatian, Lithuanian Studies: Albanian | Message 8 of 10 02 November 2014 at 9:31pm | IP Logged |
|
I wrote a new blog post about n-gram frequencies and how I use them to find
out more about Albanian. For example, they allow me to quickly find all words of a word family sorted by frequency and then find out which prepositions are used with which form. I feel I also made quite some progress in the last days and found out the meanings of many words, mainly nouns (all added to the spreadsheet). Especially identifying the Albanian equivalents of USA, UNO, EU and the likes helped a lot to understand some structures. I can also follow some very simple sentences when I'm familiar with the context, e.g. "Switzerland is a member of X, Y and Z, but not the EU". Edited by Eginhard on 02 November 2014 at 9:38pm 3 persons have voted this message useful |
|  | 
Printable version
|
| Copyright 2024 FX Micheloud - All rights reserved No part of this website may be copied by any means without my written authorization. |
