28.06.2016 22:58

Invited Talk by Fabian Barteld: Detecting spelling variants in non-standard texts

The Language Technology Group is pleased to announce the talk of

Fabian Barteld (University Hamburg)

with the title

Detecting spelling variants in non-standard texts

The talk will take place on July 6th from 11:00-12:30 at S2|02 A126.


Recent years have seen a growing interest in applying natural language processing tools to domains that feature non-standardized spelling. These domains include very different data such as historical texts and computer-mediated communication (CMC). In computational linguistics, spelling variation in these types of texts have mainly been approached as deviations from a corresponding standard, i.e., standard language in the case of CMC and a corresponding modern standard in the case of historical texts. Consequently, techniques like spell checking, rule-based string transduction, string similarity and machine translation are used to find the standard variant for non-standard spellings. In my talk, I will present experiments with an alternative approach for dealing with spelling variation where variants of the same word form are detected in the data without the reference to a standard. Such an alternative approach allows handling spelling variation in domains that lack a standardized variant and for word forms for which no standard variant exist, e.g., emoticons in CMC and extinct words in historical texts.


Fabian Barteld has studied German philology, media science, mathematics and computer science. He currently works at the University of Hamburg in the project "Referenzkorpus Mittelniederdeutsch/ Niederrheinisch (1200-1650)" (Reference corpus Middle Low German/ Low Rhenish; http://referenzkorpus-mnd-nrh.de/) and does his PhD research on spelling variation.

Related news:

Allgemeine News

A A A | Drucken Print | Impressum Impressum | Sitemap Sitemap | Suche Search | Kontakt Contact | Webseitenanalyse: Mehr Informationen
zum Seitenanfangzum Seitenanfang