Semantic sentiment analysis in social streams / Hassan Saif.

Format:

Book

Thesis/Dissertation

Author/Creator:

Saif, Hassan, author.

Series:

Studies on the Semantic Web ; Volume 030.

Studies on the Semantic Web, 2215-0870 ; Volume 030

Language:

English

Subjects (All):

Semantic computing.

Social media.

Physical Description:

1 online resource (286 pages) : illustrations, tables.

Edition:

1st ed.

Place of Publication:

Amsterdam, Netherlands ; Berlin, Germany : IOS Press : Akademische Verlagsgesellschaft, 2017.

Summary:

Microblogs and social media platforms are now considered among the most popular forms of online communication.Through a platform like Twitter, much information reflecting people's opinions and attitudes is published and shared among users on a daily basis.

Contents:

Title Page

Dedication

Statement

Abstract

Acknowledgements

Contents

List of Figures

List of Tables

1 Introduction

1.1 Motivation

1.1.1 Sentiment Analysis of Twitter: Gaps and Challenges

1.1.2 From Affect Words to Words' Semantics

1.2 Research Questions, Hypotheses and Contributions

1.3 Thesis Methodology and Outline

1.4 Publications

Background

2 Literature Review

2.1 Background

2.1.1 Fundamentals

2.1.2 A Note on Terminology

2.2 Sentiment Analysis of Twitter

2.2.1 Traditional Sentiment Analysis Approaches

2.2.1.1 The Machine Learning Approach

2.2.1.2 The Lexicon-based Approach

2.2.1.3 The Hybrid Approach

2.2.1.4 Discussion

2.3 Semantic Sentiment Analysis

2.3.1 Contextual Semantics

2.3.2 Conceptual Semantics

2.4 Summary and Discussion

2.4.1 Discussion

Semantic Sentiment Analysis of Twitter

3 Contextual Semantics for Sentiment Analysis of Twitter

3.1 Introduction

3.2 The SentiCircle Representation of Words' Semantics

3.2.1 Overview

3.2.2 SentiCircle Construction Pipeline

3.2.2.1 Term Indexing

3.2.2.2 Context Vector Generation

3.2.2.3 SentiCircle Generation

3.2.2.4 Senti-Median: The Overall Contextual Sentiment Value

3.3 SentiCircles for Sentiment Analysis

3.3.1 Entity-level Sentiment Detection

3.3.2 Tweet-level Sentiment Detection

3.3.2.1 The Median Method

3.3.2.2 The Pivot Method

3.3.2.3 The Pivot-Hybrid Method

3.3.3 Evaluation Setup

3.3.3.1 Datasets

3.3.3.2 Sentiment Lexicons

3.3.3.3 Baselines

3.3.3.4 Thresholds and Parameters Tuning

3.3.4 Evaluation Results

3.3.4.1 Entity-Level Sentiment Detection

3.3.4.2 Tweet-Level Sentiment Detection

3.3.4.3 Impact on Words' Sentiment

3.4 SentiCircles for Adapting Sentiment Lexicons.

3.4.1 Evaluating SentiStrength on the Adapted Thelwall-Lexicon

3.5 Runtime Analysis

3.6 Discussion

3.7 Summary

4 Conceptual Semantics for Sentiment Analysis of Twitter

4.1 Introduction

4.2 Conceptual Semantics for Supervised Sentiment Analysis

4.2.1 Extracting Conceptual Semantics

4.2.2 Conceptual Semantics Incorporation

4.2.3 Evaluation Setup

4.2.3.1 Datasets

4.2.3.2 Semantic Concepts Extraction

4.2.3.3 Baselines

4.2.4 Evaluation Results

4.2.4.1 Results on Incorporating Semantic Features

4.2.4.2 Comparison of Results

4.3 Conceptual Semantics for Lexicon-based Sentiment Analysis

4.3.1 Enriching SentiCircles with Conceptual Semantics

4.3.2 Evaluation Results

4.4 Discussion

4.5 Summary

5 Semantic Patterns for Sentiment Analysis of Twitter

5.1 Introduction

5.2 Related Work

5.3 Semantic Sentiment Patterns of Words

5.3.1 Syntactical Preprocessing

5.3.2 Capturing Contextual Semantics and Sentiment of Words

5.3.3 Extracting Patterns from SentiCircles

5.4 Evaluation Setup

5.4.1 Tweet-Level Evaluation Setup

5.4.2 Entity-Level Evaluation Setup

5.4.3 Evaluation Baselines

5.4.4 Number of SS-Patterns in Data

5.5 Evaluation Results

5.5.1 Results of Tweet-Level Sentiment Classification

5.5.2 Results of Entity-Level Sentiment Classification

5.6 Within-Pattern Sentiment Consistency

5.6.1 Sentiment Consistency vs. Sentiment Dispersion

5.7 Discussion

5.8 Summary

Analysis Study

6 Stopword Removal for Twitter Sentiment Analysis

6.1 Introduction

6.2 Stopword Analysis Set-Up

6.2.1 Datasets

6.2.2 Stopword removal methods

6.2.2.1 The Classic Method

6.2.2.2 Methods based on Zipf's Law (Z-Methods)

6.2.2.3 Term Based Random Sampling (TBRS)

6.2.2.4 The Mutual Information Method (MI)

6.2.3 Twitter Sentiment Classifiers.

6.3 Evaluation Results

6.3.1 Classification Performance

6.3.2 Feature Space

6.3.3 Data Sparsity

6.3.4 The Ideal Stoplist

6.4 Discussion

6.5 Summary

Conclusion

7 Discussion and Future Work

7.1 Discussion

7.1.1 Extracting Words' Semantics

7.1.1.1 Extracting Contextual Semantics

7.1.1.2 Extracting Conceptual Semantics

7.1.1.3 Extracting Stopwords

7.1.2 Incorporating Words' Semantics in Sentiment Analysis

7.1.2.1 Incorporating Semantics into Lexicon-based Approaches

7.1.2.2 Incorporating Semantics into Machine Learning Approaches

7.1.2.3 Words' Semantics for Adapting Sentiment Lexicons

7.1.3 Assessment and Results

7.2 Future Work

8 Conclusion

8.1 Contextual Semantics for Sentiment Analysis of Twitter

8.2 Conceptual Semantics for Sentiment Analysis of Twitter

8.3 Semantic Patterns for Sentiment Analysis of Twitter

8.4 Analysis on Stopword Removal Methods for Sentiment Analysis of Twitter

Appendix

A Evaluation Datasets for Twitter Sentiment Analysis

B Annotation Booklet for the STS-Gold Dataset

Bibliography.

Notes:

Ph.D. The Open University--Knowledge Media Institute 2015.

Includes bibliographical references.

Description based on online resource; title from PDF title page (ebrary, viewed August 8, 2017).

Description based on publisher supplied metadata and other sources.

ISBN:

1-61499-751-9

OCLC:

993432979