Affective Computing and Sentiment Analysis: Emotion, Metaphor and Terminology

ISBN-13: 9789400717565 / Angielski / Twarda / 2011 / 150 str.

Khurshid Ahmad

Opis

Sentiment analysis will play a key role in intelligent information extraction from text corpora ranging from telephone transcripts recorded surreptitiously to opinions expressed openly about consumer goods, books and films for example. In between, we have sentiments expressed covertly or overtly in order to discover the price of financial instruments, to influence a voting public or numerous other contexts. Sentiment analysis may yet turn out to be as big a challenge as machine translation, corpus-based lexicography, and information extraction were in their time. This is a truly multi-disciplinary effort which draws on and can inform research in psychology, artificial intelligence, knowledge management, human computer interaction, affective computing and has an impact on such diverse areas as film studies, homeland security, translation studies and others.

Spis treści

Introduction: Affect Computing and Sentiment Analysis . References .- 1. Understanding Metaphors: The Paradox of Unlike Things Compared by Sam Glucksberg . 1.1 Introduction . 1.2 The Metaphor Paraphrased Problem and the Priority of the Literal . 1.3 Understanding Metaphors: Comparison or Categorization? . 1.4 How Novel Categories Can be Named: Dual Reference . 1.5 Understanding Metaphors and Similes . 1.6 The Metaphor Paraphrase Problem Revisited . 1.7 Comparison Versus Categorization Revisited . 1.8 Conclusions . References .- 2. Metaphor as Resources for the Conceptualisation and Expression of Emotion by Andrew Goatly . 2.1 Background . 2.2 Metaphorical Conceptualisation of Emotions in English . 2.2.1 Conceptualisation of Emotion . 2.2.2 Description and Expression of Emotion . 2.3 Contribution of English Metaphor Themes to the Expression of Emotion . 2.3.2 Metalude Data for Evaluation . 2.3.2 Evaluative Transfer . 2.3.3 Evaluation Dependent on Larger Schemata . 2.3.4 Ideology and Evaluation . 2.3.5 The Role of Multivalency and Opposition in Metaphor Themes . 2.4 Conclusion . References .- 3. The Deep Lexical Semantics of Emotions by Jerry R. Hobbs and Andrew Gordon . 3.1 Introduction . 3.2 Identifying the Core Emotion Words . 3.3 Filling out the Lexicon of Emotion . 3.4 Some Core Theories . 3.5 The Theory and Lexical Semantics of Emotion . 3.6 Summary . References .- 4. Genericity and Metaphoricity Both Involve Sense Modulation by Carl Vogel . 4.1 Background . 4.2 Dynamics of First-order Information . 4.2.1 Some Intuitions About Revision . 4.2.2 A Formal Model of First-order Belief Revision . 4.2.3 First-order Belief Revision Adapted to Sense Extension . 4.3 Ramifications for Metaphoricity . 4.4 Metaphoricity and Genericity . 4.5 Particulars of the Class-Inclusion Framework . 4.6 Final Remarks . References .- 5. Affect Transfer by Metaphor for an Intelligent Conversational Agent by Alan Wallington, Rodrigo Agerri, John Barnden, Mark Lee and Tim Rumbell . 5.1 Introduction . 5.2 Affect via Metaphor in an ICA . 5.3 Metaphor Processing . 5.3.1 The Recognition Component . 5.3.2 The Analysis Component . 5.4 Examples of the Course of Processing . 5.4.1 You Piglet . 5.4.2 Lisa is an Angel . 5.4.3 Mayid is a Rock . 5.4.4 Other Examples . 5.5 Results . 5.6 Conclusions and Further Work . References .- 6. Detecting Uncertainty in Spoken Dialogues: An Explorative Research for the Automatic Detection of Speaker Uncertainty by Using Prosodic Markers by Jeroen Dral, Dirk Heylen and Rieks op den Akker . 6.1 Introduction . 6.2 Related Work 6.2.1 Defining (un)certainty . 6.2.2 Linguistic Pointers to Uncertainty . 6.2.3 Prosodic Markers of Uncertainty . 6.3 Problem Statement . 6. 4 Data Selection . 6.4.1 Selection of Meetings . 6.4.2 Data Preparation and Selection . 6.4.3 Statistical Analysis . 6.5 Experimentation . 6.5.1 Hedges –vs-No Hedges . 6.5.2 Uncertain Hedges-vs-Certain Hedges . 6.5.3 Distribution of Hedges over Dialog Acts . 6.6 Conclusions . References .- 7. Metaphors and Metaphor-like Processes Across Languages: Notes on English and Italian Language of Economics by Maria Teresa Musacchio . 7.1 Introduction . 7.2 Corpus and Method . 7.2.1 Corpus . 7.2.2 Method . 7.3 Analysis . 7.3.1 Constitutive Metaphors . 7.3.2 Pedagogic Metaphors . 7.3.3 Universal vs Culture-specific Metaphors . 7.4 Conclusion . References .- 8. The ‘Return’ and ‘Volatility’ of Sentiments: An Attempt to Quantify the Behaviour of the Markets? by Khurshid Ahmad . 8.1 Introduction . 8.2 Metaphors of ‘Return’ and ‘Volatility’ . 8.3 The Roots of Computational Sentiment Analysis . 8.4 A Corpus-based Study of Sentiments, Terminology and Ontology over Time . 8.4.1 Corpus Preparation and Composition . 8.4.2 Candidate Terminology and Ontology . 8.4.3 Historical Volatility in our Corpus . 8.5 Afterword . References .- 9 Sentiment Analysis Using Automatically Labelled Financial News Items by Michel Généreux, Thierry Poibeau and Moshe Koppel . 9.1 Introduction .  9.2 Data and Method . 9.2.1 Training and Testing Corpus . 9.2.2 Feature Types . 9.2.3 Feature Selection and Counting Methods . 9.2.4 News Items and Stock Prices Correlation . 9.2.5 Feature Selection and Semantic Relatedness of Documents . 9.3 Results . 9.3.1 Horizon Effect . 9.3.2 Polarity Effect . 9.3.3 Range Effect . 9.3.4 Effect of Adding a Neutral Class on Non-Contemporaneous Prices: One- and Two Days Ahead . 9.3.5 Conflating Two Classes . 9.3.6 Positive and Negative Features . 9.4 Discussion . 9.5 Conclusion and Future Work . References .- 10 Co-Word Analysis for Assessing Consumer Associations: A Case Study in Market Research by Thorsten Teichert, Gerhard Heyer, Katja Schöntag and Patrick Mairif . 10.1 Introduction . 10.2 Conceptual Background . 10.2. 1 Consumer Associations and Mental Processing . 10.2.2 Drawbacks of Manual Data Analysis . 10.2.3 Requirements for Automated Co-Word Analysis . 10.3 Technique and Implementation . 10 3. 1 Import of Text Sources . 10.3.2 Processing of Text . 10.3.3 Graph Creation and Clustering . 10.4 Exemplary Case Study . 10.5 Conclusion and Outlook . References .- 11 Automating Opinion Analysis in Film Reviews: The Case of Statistic Versus Linguistic Approach by Damien Poirier, Cécile Bothorel, Émilie Guimier De Neef, and Marc Boullé . 11.1 Introduction . 11.2 Related Work . 11.2.1 Machine Learning for Opinion Analysis . 11.2.2 Linguistic Methods of Opinion Analysis . 11.2.2 Linguistic Methods of Opinion Analysis . 11.3 Linguistic and Machine Learning Methods: A Comparative Study . 11.3.1 Linguistic Approach . 11.3.2 Machine Learning Approach . 11.4 Conclusion and Prospects . References .- Afterword: ‘The Fire Sermon’ by Yorick Wilks . References .- Name Index .- Subject Index

Blurb

This volume maps the watershed areas between two 'holy grails' of computer science: the identification and interpretation of affect – including sentiment and mood. The expression of sentiment and mood involves the use of metaphors, especially in emotive situations. Affect computing is rooted in hermeneutics, philosophy, political science and sociology, and is now a key area of research in computer science. The 24/7 news sites and blogs facilitate the expression and shaping of opinion locally and globally.  Sentiment analysis, based on text and data mining, is being used in the looking at news and blogs for purposes as diverse as: brand management, film reviews, financial market analysis and prediction, homeland security. There are systems that learn how sentiments are articulated.

This work draws on, and informs, research in fields as varied as artificial intelligence, especially reasoning and machine learning, corpus-based information extraction, linguistics, and psychology.