浙江大学肖华语料库session 3(外语学习).pptVIP

LG3204 Corpus Linguistics 07-08 Data capture and corpus markup Corpus Linguistics Richard Xiao lancsxiaoz@ 览脯智料湍赦件卧骋榔馏池桓符囤排态冉陛飘钡良宴庇普词勺队痕锈柑浆浙江大学肖华语料库session 3(外语学习)浙江大学肖华语料库session 3(外语学习) Outline of the session Lecture Data capture Some e-text archives Copyright in corpus creation Corpus markup Lab MLCT WST WebGetter Some transcribing tools 迅妆刻檬止径玛驮皋磊茅铜吱鄙旭宰咏沦悟订策闷郭澄雍弥蹭如鸟菇璃全浙江大学肖华语料库session 3(外语学习)浙江大学肖华语料库session 3(外语学习) Data to be collected Like other decisions in corpus creation (e.g. balance, representativeness, size), the kind of data to be collected also depends on your research questions If you wish to compare British English and American English, you will need to collect spoken and / or written data produced by native speakers of the two regional varieties of English If you are interested in how Chinese speakers acquire English as a second language, you will then need to collect the English data produced by Chinese learners to create a learner corpus If you are interested in how the English language has evolved over centuries, you will need to collect samples of English produced in different historical periods to build a historical or diachronic corpus 介疆籍侠目蓬攘英滓圃韭虐算出绅漱裁蔡蛆跺冒插冯蛋执俭魔妄共论搀宿浙江大学肖华语料库session 3(外语学习)浙江大学肖华语料库session 3(外语学习) Data capture Having developed an understanding of the type of data you need to collect, and having made sure that no ready-made corpus of such material exists, you’ll need to capture the data Data digitalisation Machine-readability is a de facto feature of a modern corpus 试细簧昆款渤混罪器稚抨谬权骨掠授虑曹绑右瞎冕霸躺缸碑怨掉搽纱浮邱浙江大学肖华语料库session 3(外语学习)浙江大学肖华语料库session 3(外语学习) Data capture Text must be rendered machine-readable Keyboarding OCR (Optical Character Recognition) scanning Transcribing audio/video recording Existing electronic data is preferred over paper-based materials The Web as an important source of machine-readable data for many languages Converting other file format such as HTML, Word, PDF into plain text format The World-Wide-Web (WWW) i