Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


SmriteDua2.pdf (428.36 KB)

ETD Abstract Container

Abstract Header

Introducing Semantic Role Labels and Enhancing Dependency Parsing to Compute Politeness in Natural Language

Dua, Smrite
http://rave.ohiolink.edu/etdc/view?acc_num=osu1430876809

Abstract Details

2015, Master of Science, Ohio State University, Computer Science and Engineering.
Politeness is a crucial aspect of natural language and affects the way we communicate with others. Computing politeness in natural language can help analyze different social factors and derive relationships amongst them. Previous work to compute politeness in data primarily focuses on exploiting lexical information in the form of unigrams and specific terms and phrases (Danescu-Niculescu-Mizil et al., 2013). The main goal of this work is to analyze how semantic information can be combined with basic lexical information to compute politeness. We build a politeness classifier that deploys basic lexical information using unigrams and combines it with semantic knowledge in the form of semantic role labels and dependency labels in our data. Experimental results for in-domain setup show improved performance of the politeness framework with the added semantic information and support our objective of evaluating politeness in a generic way by including language semantics than just certain lexical terms or phrases relating to politeness. Additionally, we deploy classification rules in our model, as we believe rules can associate politeness more accurately and couple semantic and lexical insights in language at the same time. Our in-domain experiments verify the rationale behind including rules in the classifier model and support the hypothesis of politeness determination in a broad way. Finally, we do more analysis on the top ranked semantic role label patterns with respect to their precision and coverage in the data.
Eric Fosler-Lussier (Advisor)
Alan Ritter (Committee Member)
63 p.
Computer Science
Natural language processing, Computational social science

Recommended Citations

Citations

  • Dua, S. (2015). Introducing Semantic Role Labels and Enhancing Dependency Parsing to Compute Politeness in Natural Language [Master's thesis, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1430876809

    APA Style (7th edition)

  • Dua, Smrite. Introducing Semantic Role Labels and Enhancing Dependency Parsing to Compute Politeness in Natural Language. 2015. Ohio State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1430876809.

    MLA Style (8th edition)

  • Dua, Smrite. "Introducing Semantic Role Labels and Enhancing Dependency Parsing to Compute Politeness in Natural Language." Master's thesis, Ohio State University, 2015. http://rave.ohiolink.edu/etdc/view?acc_num=osu1430876809

    Chicago Manual of Style (17th edition)

osu1430876809
856
© 2015, some rights reserved.Creative Commons License Introducing Semantic Role Labels and Enhancing Dependency Parsing to Compute Politeness in Natural Language by Smrite Dua is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by The Ohio State University and OhioLINK.