Syllabus | Assignments | People | Corpora & Tutorials | Readings | Offical LSA course page
Corpora and Tutorials
1. Logging onto the corpus server
To do your corpus work, you'll have to log onto the corpus server via SSH. Mac users and those of you using the Windows computers in the computer facilities on campus are all set. For those of you bringing private laptops that run Windows and don't already have an ssh/scp program, please download and install one, e.g. PuTTY (http://www.chiark.greenend.org.uk/~sgtatham/putty/) or OpenSSH (http://sourceforge.net/projects/sshwindows/files/OpenSSH%20for%20Windows%20-%20Release/setupssh381-20040709.zip).
If your username is lsa1 to lsa30 log onto the corpus server:
ssh <username>@174.129.5.193
If your username is lsa31 to lsa99 log onto the corpus server:
ssh <username>@174.129.205.212
Usernames and passwords will be distributed in the first class meeting.
...and if you don't know what any of this means, don't panic - we'll have a tutorial during the first week of class where we'll explain how to log on to the server and use basic Unix commands.
2. Unix Tutorial
If you don't have any experience with Unix (or if you want to refresh your memory), there will be a tutorial on Tuesday, 7/7, at 6:30pm in 212 Wheeler Hall. We'll go through basic commands that you need to navigate a Unix system. All corpus work will be done on a Unix system, so if you don't feel comfortable with Unix, please come to the tutorial! If you have a laptop, bring it along. If you don't, there will also be computers available in the room.
You can download today's Unix/vi/tgrep2 cheatsheet here: cheatsheet.txt Note that there is a Part III on regular expressions that we didn't address. Don't worry about this.
3. Available Corpora
The following corpora are available on the server for TGrep2 searches:
- Arabic Treebank (/corpora/TGrep2able/arabic-collapsed.t2c.gz)
- British National Corpus (full) (/corpora/TGrep2able/BNC.parsed.t2c.gz)
- British National Corpus (spoken) (/corpora/TGrep2able/BNC_spoken.parsed.t2c.gz)
- British National Corpus (written) (/corpora/TGrep2able/BNC_written.parsed.t2c.gz)
- Brown Corpus (/corpora/TGrep2able/brown.t2c.gz)
- Chinese Trebank (/corpora/TGrep2able/chtb6.t2c.gz)
- International Corpus of English (/corpora/TGrep2able/icegb.t2c.gz)
- NEGRA (/corpora/TGrep2able/negra.t2c.gz) -- written German
- Switchboard Corpus (/corpora/TGrep2able/sw.backtrans.convid_020607.t2c.gz)
- TIGER (/corpora/TGrep2able/tiger.t2c.gz) -- written German
- Wall Street Journal (/corpora/TGrep2able/wsj_mrg.t2c.gz)
- York-Toronto-Helsinki Parsed Corpus of Old English Prose (/corpora/TGrep2able/ycoe.t2c.gz)
4. TGrep2 and the TGrep2 Database Tools (TDT)
Here is the TGrep2 User Manual with information on how to run TGrep2, Tgrep2 options and pattern syntax, how to create MACRO files, etc. TGrep2Manual.pdf
For tutorial-style introductions to TGrep2, try these: http://www.bcs.rochester.edu/people/fjaeger/teaching/tutorials/TGrep2/LabSyntax-Tutorial.html and http://www.stanford.edu/dept/linguistics/corpora/cas-tut-tgrep.html
For preliminary documentation on the TDT Tools, see the TDT page
For the Penn Treebank Bracketing Conventions: http://bulba.sdsu.edu/jeanette/thesis/PennTags.html
For the Switchboard-specific bracketing conventions: swbd_bracketing.pdf
5. Other search software
The Corpus Query Processor (CQP) - a tutorial: cqp_tutorial.pdf
6. R Tutorials
This guide is a brief schematic introduction to computing in the R language. Basic statistical concepts such as variables, descriptive statistics, scales, reasoning, hypothesis testing and power analysis are defined and explained. StatsNotes1.pdf, StatsNotes2.pdf.
Also, consider out lab-internal stats tutorials with R-scripts, reading suggestions, etc. There also are a couple of posts on our HLP lab blog about visualization, model fitting, common issues, snippets of R code, etc. The easiest way to find what you're looking for is to just enter it as a search term on the blog page linked above.
Finally, if you decide to make R your choice for data analysis, we recommend that you enroll to the R-lang email list. It's low-traffic, and directed at language researchers. The perfect place to ask questions.
There are 7 attachment(s) stored for this page.
Syllabus | Assignments | People | Corpora & Tutorials | Readings | Offical LSA course page