PMI-cool at SemEval-2016 task 3

Experiments with PMI and goodness polarity lexicons for community question answering

Daniel Balchev, Yasen Kiprov, Ivan Koychev, Preslav Nakov

Research output: Chapter in Book/Report/Conference proceedingConference contribution

8 Citations (Scopus)

Abstract

We describe our submission to SemEval-2016 Task 3 on Community Question Answering. We participated in subtask A, which asks to rerank the comments from the thread for a given forum question from good to bad. Our approach focuses on the generation and use of goodness polarity lexicons, similarly to the sentiment polarity lexicons, which are very popular in sentiment analysis. In particular, we use a combination of bootstrapping and pointwise mutual information to estimate the strength of association between a word (from a large unannotated set of question-answer threads) and the class of good/bad comments. We then use various features based on these lexicons to train a regression model, whose predictions we use to induce the final comment ranking. While our system was not very strong as it lacked important features, our lexicons contributed to the strong performance of another top-performing system.

Original languageEnglish
Title of host publicationSemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings
PublisherAssociation for Computational Linguistics (ACL)
Pages844-850
Number of pages7
ISBN (Electronic)9781941643952
Publication statusPublished - 1 Jan 2016
Event10th International Workshop on Semantic Evaluation, SemEval 2016 - San Diego, United States
Duration: 16 Jun 201617 Jun 2016

Other

Other10th International Workshop on Semantic Evaluation, SemEval 2016
CountryUnited States
CitySan Diego
Period16/6/1617/6/16

Fingerprint

Question Answering
Polarity
Thread
Sentiment Analysis
Bootstrapping
Mutual Information
Large Set
Experiment
Regression Model
Ranking
Experiments
Prediction
Estimate
Community
Class

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computational Theory and Mathematics
  • Computer Science Applications

Cite this

Balchev, D., Kiprov, Y., Koychev, I., & Nakov, P. (2016). PMI-cool at SemEval-2016 task 3: Experiments with PMI and goodness polarity lexicons for community question answering. In SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings (pp. 844-850). Association for Computational Linguistics (ACL).

PMI-cool at SemEval-2016 task 3 : Experiments with PMI and goodness polarity lexicons for community question answering. / Balchev, Daniel; Kiprov, Yasen; Koychev, Ivan; Nakov, Preslav.

SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings. Association for Computational Linguistics (ACL), 2016. p. 844-850.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Balchev, D, Kiprov, Y, Koychev, I & Nakov, P 2016, PMI-cool at SemEval-2016 task 3: Experiments with PMI and goodness polarity lexicons for community question answering. in SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings. Association for Computational Linguistics (ACL), pp. 844-850, 10th International Workshop on Semantic Evaluation, SemEval 2016, San Diego, United States, 16/6/16.
Balchev D, Kiprov Y, Koychev I, Nakov P. PMI-cool at SemEval-2016 task 3: Experiments with PMI and goodness polarity lexicons for community question answering. In SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings. Association for Computational Linguistics (ACL). 2016. p. 844-850
Balchev, Daniel ; Kiprov, Yasen ; Koychev, Ivan ; Nakov, Preslav. / PMI-cool at SemEval-2016 task 3 : Experiments with PMI and goodness polarity lexicons for community question answering. SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings. Association for Computational Linguistics (ACL), 2016. pp. 844-850
@inproceedings{13019d1195fe4ad1b3b8c056e33162fa,
title = "PMI-cool at SemEval-2016 task 3: Experiments with PMI and goodness polarity lexicons for community question answering",
abstract = "We describe our submission to SemEval-2016 Task 3 on Community Question Answering. We participated in subtask A, which asks to rerank the comments from the thread for a given forum question from good to bad. Our approach focuses on the generation and use of goodness polarity lexicons, similarly to the sentiment polarity lexicons, which are very popular in sentiment analysis. In particular, we use a combination of bootstrapping and pointwise mutual information to estimate the strength of association between a word (from a large unannotated set of question-answer threads) and the class of good/bad comments. We then use various features based on these lexicons to train a regression model, whose predictions we use to induce the final comment ranking. While our system was not very strong as it lacked important features, our lexicons contributed to the strong performance of another top-performing system.",
author = "Daniel Balchev and Yasen Kiprov and Ivan Koychev and Preslav Nakov",
year = "2016",
month = "1",
day = "1",
language = "English",
pages = "844--850",
booktitle = "SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings",
publisher = "Association for Computational Linguistics (ACL)",

}

TY - GEN

T1 - PMI-cool at SemEval-2016 task 3

T2 - Experiments with PMI and goodness polarity lexicons for community question answering

AU - Balchev, Daniel

AU - Kiprov, Yasen

AU - Koychev, Ivan

AU - Nakov, Preslav

PY - 2016/1/1

Y1 - 2016/1/1

N2 - We describe our submission to SemEval-2016 Task 3 on Community Question Answering. We participated in subtask A, which asks to rerank the comments from the thread for a given forum question from good to bad. Our approach focuses on the generation and use of goodness polarity lexicons, similarly to the sentiment polarity lexicons, which are very popular in sentiment analysis. In particular, we use a combination of bootstrapping and pointwise mutual information to estimate the strength of association between a word (from a large unannotated set of question-answer threads) and the class of good/bad comments. We then use various features based on these lexicons to train a regression model, whose predictions we use to induce the final comment ranking. While our system was not very strong as it lacked important features, our lexicons contributed to the strong performance of another top-performing system.

AB - We describe our submission to SemEval-2016 Task 3 on Community Question Answering. We participated in subtask A, which asks to rerank the comments from the thread for a given forum question from good to bad. Our approach focuses on the generation and use of goodness polarity lexicons, similarly to the sentiment polarity lexicons, which are very popular in sentiment analysis. In particular, we use a combination of bootstrapping and pointwise mutual information to estimate the strength of association between a word (from a large unannotated set of question-answer threads) and the class of good/bad comments. We then use various features based on these lexicons to train a regression model, whose predictions we use to induce the final comment ranking. While our system was not very strong as it lacked important features, our lexicons contributed to the strong performance of another top-performing system.

UR - http://www.scopus.com/inward/record.url?scp=85035754662&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85035754662&partnerID=8YFLogxK

M3 - Conference contribution

SP - 844

EP - 850

BT - SemEval 2016 - 10th International Workshop on Semantic Evaluation, Proceedings

PB - Association for Computational Linguistics (ACL)

ER -