Online German Language Corpus, UCREL Summer School

At the University of Leipzig a German Language corpus is available (Projekt Deutscher Wortschatz). The database can be queried from different programming languages and access is also possible via a web service. Requests can ask for co-occurrences of words, base forms, about words that often occur to the right and to the left of the word, word frequency, synonyms and much more. If you develop text input systems this may be a very useful resource, see the web services overview page (with links to downloads), the list of web-service-requests offered or have a look at some php-examples.

You can try the service interactively at http://wortschatz.uni-leipzig.de/abfrage/. See the pictures for an example query on the term Internet. They also feature a German-English dictionary.

Since I shared and office at Lancaster University with Paul Rayson from UCREL (University Centre for Computer Corpus Research on Language) I find corpus linguistics an interesting topic. By the way UCREL runs a Summer School in Corpus Linguistics from 13 to 15 July 2011 - would love to go there...

Male (88%), writing like Oscar Wilde (35%)

Looking into Paul Rayson's blog and discovered an interesting link: http://www.genderanalyzer.com. It is a web form where you can put in an URL and you get an estimate whether the author of this text is male or female. For me it worked great ;-) It says that the text I wrote in my blog is with 88% written by a male. I tried it with a few more of my pages and it worked. Then I looked at some pages of some of my female colleagues and to my surprise it seems they do not write their web pages by themselves (as the program indicated 95% male writer) - they probably all have a hidden male assistant ;-)

While I was in Lancaster I shared for most of the time an office with Paul. During this time I learned a lot of interesting things about corpus linguistics and phenomena in language in general - just by sharing the office. One fact at that at the time was surprising to me is that if you take 6 words from an arbitrary text in the exact order as they appear in the text and you search on the web for the exact phrase it is likely that you will only find this text. How many hits do you get for phrase "I was at Trinity College reading" in google? Try it out ;-) [to students: that is why not getting caught when you plagiarize is really hard]

From http://www.genderanalyzer.com I came to http://www.ofaust.com and to my great surprise I write like Oscar Wilde (35%) and Friedrich Nietzsche (30%). Thinking of social networks (and in particular the use of languages within closed groups) such technologies could become an interesting enabling technology for novel applications. Perhaps I should visit Paul again in Lancaster…

PS: and I nearly forgot I am a thinker / INTJ - The Scientists (according to http://www.typealyzer.com/)

PPS (2008-11-17): a further URL contrinuted from my collegues on the gender topic: http://www.mikeonads.com/2008/07/13/using-your-browser-url-history-estimate-gender/