logging in or signing up Multiword NN Expression aSGuest403 Download Post to : URL : Related Presentations : Share Add to Flag Embed Email Send to Blogs and Networks Add to Channel Uploaded from authorPOINT lite Insert YouTube videos in PowerPont slides with aS Desktop Copy embed code: (To copy code, click on the text box) Embed: URL: Thumbnail: WordPress Embed Customize Embed The presentation is successfully added In Your Favorites. Views: 33 Category: Others/ Misc License: All Rights Reserved Like it (0) Dislike it (0) Added: October 02, 2008 This Presentation is Public Favorites: 0 Presentation Description No description available. Comments Posting comment... Premium member Presentation Transcript Multiword (NN) Expression : Multiword (NN) Expression Dharmvir Kumar M.tech.(1st year), CSE dept. Indian Institute of Technology, Bombay Multiword (NN) Expression : Multiword (NN) Expression In this presentation we will discuss What is MWE(MultiWord Expression) Some Observations about MWE(NN) An approach towards handling of MWE(NN) Conclusion References What is MWE : What is MWE Definition: A multiword expression is decomposable into different simplex word but its syntactic and semantic properties cannot be derived from its parts. In other words, MWEs can be defined as cohesive lexemes that cross word boundaries . some examples: frying pan, bread and butter, ad hoc, part of speech, life science, telephone box ,coffee machine, San Francisco, stand by,…. What is MWE : What is MWE Flexible and heterogeneous nature of MWE makes it endless source of parsing failure. To provide a unified account for detection of these distinct but related phenomena is a real challenge for robustness of NLP system Some Observations about MWE(NN) : Some Observations about MWE(NN) some more examples : Prime minister, platinum jewellery,Taj Mahal,Air India, traffic signal, Mahatma Gandhi etc. Now take some sentences like : Man Mohan Singh is the Prime minister of India. Gold has tradition but platinum jewellery is the new generation’s choice. Mohandas Karamchand Gandhi is known as Mahatma Gandhi. All traffic signals will be removed using over bridges. Agra is famous due to Taj Mahal. Some Observations about MWE(NN) : Some Observations about MWE(NN) Key observations are following : In any sentence ,two or more nouns cannot occur without any prepositions or relation specifier. If two or more nouns occurs continuously, then there is a chance of multiword expression or there may be errors. An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) For the purpose of detection of MWEs and acquisition in grammar lexicons. We will go through following steps one by one. Take a well prepared corpus and Grammar which can perform tagging of POS over sentences. Perform POS tagging over sentences of corpus, one by one. An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) Take places where Noun-Noun combinations occur. These are possible MWEs. All reported combinations should be validated using World Wide Web as a big corpus. Frequencies of all permutation of candidate MWEs can be obtained using Google for exact match. for example some of observed frequency : An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) According to observation made by frequencies of all permutation ,one can take decision about validity of candidate MWEs. A valid MWE should be added in grammar lexicons. Conclusion : Conclusion The given approach of MWEs detection is capable to extend the coverage of grammar up to certain extent. If automatic acquisition of lexicons is done after validation phase then performance of coverage extension will be visible in the process itself. References : References Yi Zhang & Valia Kordani,Aline Villavicencio & Marco Idiart ,Automated Multiword Expression Prediction for Grammar Engineering. In the Proceeding of workshop on Multiword Expression ,Sydney, July 2006 (ACL) Yi Zhang & Velia Kordani.2006. Automated deep lexical acquisition for robust open text processing. In proceeding of Fifth International Conference on Language Resources and Evaluation(LREC 2006),Geona ,Italy. You do not have the permission to view this presentation. In order to view it, please contact the author of the presentation.
Multiword NN Expression aSGuest403 Download Post to : URL : Related Presentations : Share Add to Flag Embed Email Send to Blogs and Networks Add to Channel Uploaded from authorPOINT lite Insert YouTube videos in PowerPont slides with aS Desktop Copy embed code: (To copy code, click on the text box) Embed: URL: Thumbnail: WordPress Embed Customize Embed The presentation is successfully added In Your Favorites. Views: 33 Category: Others/ Misc License: All Rights Reserved Like it (0) Dislike it (0) Added: October 02, 2008 This Presentation is Public Favorites: 0 Presentation Description No description available. Comments Posting comment... Premium member Presentation Transcript Multiword (NN) Expression : Multiword (NN) Expression Dharmvir Kumar M.tech.(1st year), CSE dept. Indian Institute of Technology, Bombay Multiword (NN) Expression : Multiword (NN) Expression In this presentation we will discuss What is MWE(MultiWord Expression) Some Observations about MWE(NN) An approach towards handling of MWE(NN) Conclusion References What is MWE : What is MWE Definition: A multiword expression is decomposable into different simplex word but its syntactic and semantic properties cannot be derived from its parts. In other words, MWEs can be defined as cohesive lexemes that cross word boundaries . some examples: frying pan, bread and butter, ad hoc, part of speech, life science, telephone box ,coffee machine, San Francisco, stand by,…. What is MWE : What is MWE Flexible and heterogeneous nature of MWE makes it endless source of parsing failure. To provide a unified account for detection of these distinct but related phenomena is a real challenge for robustness of NLP system Some Observations about MWE(NN) : Some Observations about MWE(NN) some more examples : Prime minister, platinum jewellery,Taj Mahal,Air India, traffic signal, Mahatma Gandhi etc. Now take some sentences like : Man Mohan Singh is the Prime minister of India. Gold has tradition but platinum jewellery is the new generation’s choice. Mohandas Karamchand Gandhi is known as Mahatma Gandhi. All traffic signals will be removed using over bridges. Agra is famous due to Taj Mahal. Some Observations about MWE(NN) : Some Observations about MWE(NN) Key observations are following : In any sentence ,two or more nouns cannot occur without any prepositions or relation specifier. If two or more nouns occurs continuously, then there is a chance of multiword expression or there may be errors. An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) For the purpose of detection of MWEs and acquisition in grammar lexicons. We will go through following steps one by one. Take a well prepared corpus and Grammar which can perform tagging of POS over sentences. Perform POS tagging over sentences of corpus, one by one. An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) Take places where Noun-Noun combinations occur. These are possible MWEs. All reported combinations should be validated using World Wide Web as a big corpus. Frequencies of all permutation of candidate MWEs can be obtained using Google for exact match. for example some of observed frequency : An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) An approach towards handling of MWE(NN) : An approach towards handling of MWE(NN) According to observation made by frequencies of all permutation ,one can take decision about validity of candidate MWEs. A valid MWE should be added in grammar lexicons. Conclusion : Conclusion The given approach of MWEs detection is capable to extend the coverage of grammar up to certain extent. If automatic acquisition of lexicons is done after validation phase then performance of coverage extension will be visible in the process itself. References : References Yi Zhang & Valia Kordani,Aline Villavicencio & Marco Idiart ,Automated Multiword Expression Prediction for Grammar Engineering. In the Proceeding of workshop on Multiword Expression ,Sydney, July 2006 (ACL) Yi Zhang & Velia Kordani.2006. Automated deep lexical acquisition for robust open text processing. In proceeding of Fifth International Conference on Language Resources and Evaluation(LREC 2006),Geona ,Italy.