Your Sentiment Precedes You: Using an author’s historical tweets to predict sarcasm Anupam Khattri 2, Aditya Joshi 1,3, Pushpak Bhattacharyya 1, Mark James.

Slides:



Advertisements
Similar presentations
A Comparison of Implicit and Explicit Links for Web Page Classification Dou Shen 1 Jian-Tao Sun 2 Qiang Yang 1 Zheng Chen 2 1 Department of Computer Science.
Advertisements

Dan Jurafsky Lecture 4: Sarcasm, Alzheimers, +Distributional Semantics Computational Extraction of Social and Interactional Meaning SSLST, Summer 2011.
CS460/IT632 Natural Language Processing/Language Technology for the Web Lecture 2 (06/01/06) Prof. Pushpak Bhattacharyya IIT Bombay Part of Speech (PoS)
Using the SmartPLS Software “Structural Model Assessment”
A cognitive study of subjectivity extraction in sentiment annotation Abhijit Mishra 1, Aditya Joshi 1,2,3, Pushpak Bhattacharyya 1 1 IIT Bombay, India.
Entity-Centric Topic-Oriented Opinion Summarization in Twitter Date : 2013/09/03 Author : Xinfan Meng, Furu Wei, Xiaohua, Liu, Ming Zhou, Sujian Li and.
Linking Entities in #Microposts ROMIL BANSAL, SANDEEP PANEM, PRIYA RADHAKRISHNAN, MANISH GUPTA, VASUDEVA VARMA INTERNATIONAL INSTITUTE OF INFORMATION TECHNOLOGY,
Farag Saad i-KNOW 2014 Graz- Austria,
Twitter Glossary. #: People use the hashtag symbol # before a relevant keyword or phrase (no spaces) in their Tweet to categorize those Tweets and help.
Distant Supervision for Emotion Classification in Twitter posts 1/17.
Problem Semi supervised sarcasm identification using SASI
Multiple Aspect Ranking using the Good Grief Algorithm Benjamin Snyder and Regina Barzilay at MIT Elizabeth Kierstead.
Semi Supervised Recognition of Sarcastic Sentences in Twitter and Amazon Dmitry DavidovOren TsurAri Rappoport.
Sarcasm Detection on Twitter A Behavioral Modeling Approach
Title Course opinion mining methodology for knowledge discovery, based on web social media Authors Sotirios Kontogiannis Ioannis Kazanidis Stavros Valsamidis.
University of Sheffield NLP Opinion Mining in GATE Horacio Saggion & Adam Funk.
Social network analysis of the West African Ebola Outbreak INTRODUCTION The Ebola outbreak in West Africa has drawn the World’s attention given the spread.
Sentiment Lexicon Creation from Lexical Resources BIS 2011 Bas Heerschop Erasmus School of Economics Erasmus University Rotterdam
Predicting the Semantic Orientation of Adjective Vasileios Hatzivassiloglou and Kathleen R. McKeown Presented By Yash Satsangi.
A Self Learning Universal Concept Spotter By Tomek Strzalkowski and Jin Wang Presented by Iman Sen.
Duyu Tang, Furu Wei, Nan Yang, Ming Zhou, Ting Liu, Bing Qin
Extracting Interest Tags from Twitter User Biographies Ying Ding, Jing Jiang School of Information Systems Singapore Management University AIRS 2014, Kuching,
Distributed Representations of Sentences and Documents
#title We know tweeted last summer ! Shrey Gupta & Sonali Aggarwal.
EARLY DETECTION OF TWITTER TRENDS MILAN STANOJEVIC UNIVERSITY OF BELGRADE SCHOOL OF ELECTRICAL ENGINEERING.
Forecasting with Twitter data Presented by : Thusitha Chandrapala MARTA ARIAS, ARGIMIRO ARRATIA, and RAMON XURIGUERA.
A Joint Model of Feature Mining and Sentiment Analysis for Product Review Rating Jorge Carrillo de Albornoz Laura Plaza Pablo Gervás Alberto Díaz Universidad.
1 Using R for consumer psychological research Research Analytics | Strategy & Insight September 2014.
Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification on Reviews Peter D. Turney Institute for Information Technology National.
Empirical Methods in Information Extraction Claire Cardie Appeared in AI Magazine, 18:4, Summarized by Seong-Bae Park.
1 Entity Discovery and Assignment for Opinion Mining Applications (ACM KDD 09’) Xiaowen Ding, Bing Liu, Lei Zhang Date: 09/01/09 Speaker: Hsu, Yu-Wen Advisor:
Distributional Part-of-Speech Tagging Hinrich Schütze CSLI, Ventura Hall Stanford, CA , USA NLP Applications.
2007. Software Engineering Laboratory, School of Computer Science S E Towards Answering Opinion Questions: Separating Facts from Opinions and Identifying.
Profile The METIS Approach Future Work Evaluation METIS II Architecture METIS II, the continuation of the successful assessment project METIS I, is an.
When Experts Agree: Using Non-Affiliated Experts To Rank Popular Topics Meital Aizen.
On the Issue of Combining Anaphoricity Determination and Antecedent Identification in Anaphora Resolution Ryu Iida, Kentaro Inui, Yuji Matsumoto Nara Institute.
Computer Science: A Structured Programming Approach Using C1 Objectives ❏ To be able to list and describe the six expression categories ❏ To understand.
Opinion Mining of Customer Feedback Data on the Web Presented By Dongjoo Lee, Intelligent Databases Systems Lab. 1 Dongjoo Lee School of Computer Science.
COLING 2012 Extracting and Normalizing Entity-Actions from Users’ comments Swapna Gottipati, Jing Jiang School of Information Systems, Singapore Management.
Prediction of Influencers from Word Use Chan Shing Hei.
A Scalable Machine Learning Approach for Semi-Structured Named Entity Recognition Utku Irmak(Yahoo! Labs) Reiner Kraft(Yahoo! Inc.) WWW 2010(Information.
TEXT ANALYTICS - LABS Maha Althobaiti Udo Kruschwitz Massimo Poesio.
Recommender Systems Debapriyo Majumdar Information Retrieval – Spring 2015 Indian Statistical Institute Kolkata Credits to Bing Liu (UIC) and Angshul Majumdar.
Multilingual Opinion Holder Identification Using Author and Authority Viewpoints Yohei Seki, Noriko Kando,Masaki Aono Toyohashi University of Technology.
MedKAT Medical Knowledge Analysis Tool December 2009.
Recognizing Stances in Online Debates Unsupervised opinion analysis method for debate-side classification. Mine the web to learn associations that are.
Multi-level Bootstrapping for Extracting Parallel Sentence from a Quasi-Comparable Corpus Pascale Fung and Percy Cheung Human Language Technology Center,
Iterative similarity based adaptation technique for Cross Domain text classification Under: Prof. Amitabha Mukherjee By: Narendra Roy Roll no: Group:
Finding document topics for improving topic segmentation Source: ACL2007 Authors: Olivier Ferret (18 route du Panorama, BP6) Reporter:Yong-Xiang Chen.
Learning Extraction Patterns for Subjective Expressions 2007/10/09 DataMining Lab 안민영.
Reputation Management System
From Words to Senses: A Case Study of Subjectivity Recognition Author: Fangzhong Su & Katja Markert (University of Leeds, UK) Source: COLING 2008 Reporter:
An evolutionary approach for improving the quality of automatic summaries Constantin Orasan Research Group in Computational Linguistics School of Humanities,
Semi-Supervised Recognition of Sarcastic Sentences in Twitter and Amazon -Smit Shilu.
Twitter as a Corpus for Sentiment Analysis and Opinion Mining
Project Deliverable-1 -Prof. Vincent Ng -Girish Ramachandran -Chen Chen -Jitendra Mohanty.
Twitter Hashtags RMBI4310Spring 2016 Group 14 Cheung Hiu Yan, Debbie Chow Miu Lam, Carman Tsang Wing Wah, Denise
More than words: Social network’s text mining for consumer brand sentiments Expert Systems with Applications 40 (2013) 4241–4251 Mohamed M. Mostafa Reporter.
A Sentiment-Based Approach to Twitter User Recommendation BY AJAY ABDULPUR RAJARAM NIKKAM.
Applications of Text Analytics: Drunk-texting and Sarcasm Aditya Joshi IITB-Monash Research Academy Joint work with Abhijit Mishra, Vinita Sharma, Balamurali.
Homework 3 Progress Presentation -Meet Shah. Goal Identify whether tweet is sarcastic or not.
Sentiment Analysis of Twitter Messages Using Word2Vec
Name: Sushmita Laila Khan Affiliation: Georgia Southern University
Sentiment Analysis on Interactive Conversational Agent/Chatbots
February 16, 2012 – Carmen Brenner
Sentiment Analysis Study
SOCIAL COMPUTING Homework 3 Presentation
Do humans beat computers at pattern recognition? Andra Miloiu Costina
Text Mining & Natural Language Processing
Presentation transcript:

Your Sentiment Precedes You: Using an author’s historical tweets to predict sarcasm Anupam Khattri 2, Aditya Joshi 1,3, Pushpak Bhattacharyya 1, Mark James Carman 3 1 Indian Institute of Technology Bombay, India 2 Indian Institute of Technology Kharagpur, India 3 Monash University, Australia First presented at 6 th WASSA at EMNLP, September 2015

The Key Message  Sarcasm detection must look beyond target text to identify nuanced forms of sarcasm such as hyperbole.  For a target tweet, we look up the historical tweets written by the author to detect sarcasm

Background  Maynard et al (2015) compute sentiment of a tweet and the sentiment of the tokenized hashtag. If the two do not match, then predict as sarcastic.  Joshi et al.(2015) used context incongruity within text to identify sarcasm within text

There may not always be ‘incongruity’ in the target tweet!

How we do it? A tweet is sarcastic if: 1. There is contrast in target tweet 2. There is incompatibility between sentiment in this tweet, and towards historical tweets in the past that contain the entity in the target tweet

Architecture

 Contrast of sentiments in the same tweet is a very strong indication of sarcasm.  Explicit - When words from wordlist with opposite sentiment are present in the same target tweet, then we mark it as sarcastic.  Implicit -We find these phrases by the bootstrap algorithm described in Tsur et al. (2010) Davidov et al.(2010) (based on sentiment hashtags in tweets and commonly occurring phrases) Contrast Based Predictor

Historical Tweets  We run POS- tagger on target tweet.  Get tweets from user’s timeline which have NNP sequence as in the target tweet.  These tweets are termed Historical tweets.  This way our approach extends the past work by considering sentiment contrasts beyond the target tweet.

 Historical text is taken as the users view about the subject (NNP). Why only NNP?  Sentiment of all the historical tweets in averaged to get the users views.  Sentiment is calculated using a simple rule-based engine that looks up lexicon and rules for negation and conjunctions  If the target tweet sentiment is in contrast to the historical sentiment then target tweet is classified as sarcastic else non- sarcastic. Historical Tweet Based Predictor

Integrator  We combine the results of our prediction of tweets as sarcastic and non-sarcastic from the contrast based predictor and historical tweet based predictor in three ways :-  1. OR – A tweet is sarcastic if predicted as sarcastic by at least one predictor.  2, AND – A tweet is sarcastic if predicted as sarcastic by both the predictors.  3. Relaxed AND – A tweet is sarcastic if predicted as sarcastic by both or if the historical tweets are not found then the result of contrast based predictor is considered.

Experimental Setup  Test tweets from Riloff et al. (2013): total tweets, 506- Sarcastic tweets, non-sarcastic tweets.  For contrast-based predictor: Extracted from 8000 ‘#sarcastic’ tweets. 442 phrases with implicit negative sentiment (no implicit positive sentiment phrases because we extracted phrases from sarcastic tweets  For historical sentiment-based predictor: Wordlists from Pang et al (2004) and Mohammad et al (2013)

Results (1/2) PrecisionRecallF- Score Best Reported value by Riloff et al(2013) Only Historical Tweet-based OR AND Relaxed – AND Using Pang and Lee (2004) wordlist

Results (2/2) Using Mohammad and Turney (2013) wordlist PrecisionRecallF- Score Best Reported value by Riloff et al(2013) Only Historical Tweet-based OR AND Relaxed – AND

Discussion (1/2) As the target phrases in our approach are NNP sequence we were able to predict sarcasm in the following instances:-  Tweets addressed to other users. By historical tweets we were able to find the relation between users and therefore sarcasm.  Tweets referring to companies or organizations. Ex. Target tweet - “because Fox’s is well-balanced and objective ? ”. Historical Tweet – “Fox’s World cup streaming options are terrible.”  Informal words and expressions were tagged as NNPs and thus from historical tweets we could extract the sentiment associated with it and hence predict sarcasm.

Discussion (2/2) Limitation of our approach :  The assumption that our historical tweet-based predictor makes is that the author has not been sarcastic about a target phrase in the past.  A twitter timeline may be private or twitter handle name was changed or account was deactivated.

Conclusion & Future Work  We present a approach that predicts sarcasm in a target tweet using the tweet author’s historical tweets.  We obtain the best F-score value of 0.882, when the contrast predictor uses a set of polar words from a word-emotion lexicon and phrases with implicit sentiment, and “RELAXED-AND” is used.  Our work opens a new direction to sarcasm detection: considering text written by an author in the past to identify sarcasm in a piece of text.