These are assorted research resources developed as part of projects that I have worked on.
The Jinan Chinese Learner Corpus (JCLC) is a 6 million token collection of texts written by Chinese L2 learners. Learn more about the learner corpus.
Journalistic text from closely related languages and language varieties used in the 2016 DSL Shared Task. Click here for more details and download information.
The MQ Names Corpus contains over 13k names from 5 cultural groups, which have been annotated for gender. Details and access info.
A set of Brown clusters induced from 100 million tokens of medical research abstracts. Click here to view the clusters.
They were used to extract classification features in this paper.
As part of the research on Chinese Native Language Identification, we compiled a list of 449 Chinese function words to be used as features in our model. The function word list was compiled from Chinese language teaching resources. The complete list can be accessed here.
As part the work on Norwegian Native Language Identification we used a list of 176 function words obtained from the distribution of the Apache Lucene search engine software. This list includes stop words for the Bokmål variant of the language and contains entries such as hvis (whose), ikke (not) and jeg (I). The list is available here.
This is a corpus of Dari news texts that was used to discriminate between Farsi and Dari. Learn more about the Dari corpus or click here to read the paper.
This is a corpus of Sorani Kurdish sentences that was used to discriminate between two subdialects. Learn more about the Sorani corpus or click here to read the paper.