中国政府26日式宣布控制温室气体排放的行动目标.pptVIP

Natural Language Processing LING101 December 3, 2009 Quickwrite #22 Suppose you wanted to develop a spelling detection system, which runs through a text file and detects spelling errors. One obvious thing to do would be to have your system look up each word in a dictionary: if it is not in the dictionary, it is an error. This may not always work as you can see in the following example: Your absolutely right, its not going to work How would you revise your system to handle such cases? Natural language processing The term in today’s class means computer processing of written text as opposed to speech This is a narrow interpretation of the term as it is sometimes used interchangeably with computational linguistics Example applications Spelling error detection / correction Grammar checker Search engines Machine translation Example underlying linguistic problems Morpheme/word boundary detection Part-of-speech tagging Parsing Semantic role labeling Named entity recognition Transliteration Word, phrase, sentence alignment Word boundary detection Not all writing systems mark word boundary 中国政府26日正式宣布控制温室气体排放的行动目标，到2020年单位国内生产总值(GDP)二氧化碳排放比2005年下降40%至45%。据专家估算，为达到这一目标，中国每个家庭每年要为此多支付64美元，并大大增加了中国的现代化成本 Part-of-speech (POS) tagging Tag the words with their syntactic category (POS) Dictionary look-up will not always work because of ambiguity Will will write his will in a castle Will – Proper noun will – Auxiliary will – Noun Parsing Will will write his will in a castle Will_Noun will_Aux write_verb his_Det will_Noun in_Prep a_Det castle_Noun Will can will write his will in a castle Will_Noun can_ Aux will_Aux write_verb his_Det will_Noun in_Prep a_Det castle_Noun S - NP Aux VP NP - Noun VP - VP PP VP - VP NP VP - Verb NP - Det Noun PP - Prep NP Semantic role labeling Label the words by their semantic roles with respect to the verb Will will write his will in a castle Verb: write Agent: Will Theme: his will Location: in a castle Named entity recognition Havi