From Valence to Emotions

Table of Contents

• Abstract
• List of Abbreviations
• List of Figures
• List of Tables
• 1 Introduction
• 2 Structure of Thesis
• 3 The Need of Automated Prediction Using Online Sentiments
• 4 What are the Different Prediction and Sentiment Detection Approaches and Techniques based on User-Generated-Content?
• 4.1 User Generated Content and its Technical Background
• 4.1.1 Social Media vs. Web 2.0
• 4.1.2 Online Community
• 4.1.3 Social Networking Service
• 4.1.4 Weblog
• 4.1.5 Review Site
• 4.2 Online Word-of-Mouth
• 4.2.1 Appearance of Online Word-of-Mouth
• 4.2.1.1 Scale Rating
• 4.2.1.2 Tweets
• 4.2.1.3 Review Texts
• 4.2.1.4 Blog Posts
• 4.2.2 Forms of Online Sentiments
• 4.2.2.1 Volume
• 4.2.2.2 Valence
• 4.2.2.3 Emotions
• 4.3 Sentiment Classification
• 4.3.1 Machine Learning Techniques
• 4.3.1.1 Naïve Bayes
• 4.3.1.2 Maximum Entropy
• 4.3.1.3 Support Vector Machines
• 4.3.2 Semantic Orientation Approach
• 4.3.2.1 Pointwise Mutual Information and Information Retrieval
• 4.3.2.2 Latent Semantic Analysis
• 5 How Consistent are Prediction Results Based on Online Sentiments?
• 5.1 Predictive Power of Online Sentiments
• 5.1.1 Stock Markets
• 5.1.1.1 Predictive Sources
• 5.1.1.2 Methods and Findings
• 5.1.1.3 Consistency
• 5.1.1.4 Limitations
• 5.1.2 Sales Volume
• 5.1.2.1 Predictive Sources
• 5.1.2.2 Methods and Findings
• 5.1.2.3 Consistency
• 5.1.2.4 Limitations
• 5.1.3 Box Office Revenues
• 5.1.3.1 Predictive Sources
• 5.1.3.2 Methods and Findings
• 5.1.3.3 Consistency
• 5.1.3.4 Limitations
• 6 Do Fine-Grained Sentiments Generate New Insights and Better Prediction Results Than Coarse Sentiments?
• 7 Conclusion
• 8 Managerial Implications
• Bibliography

Objectives and Key Themes

This thesis aims to provide a comprehensive overview of existing research on predicting real-world outcomes using online sentiments. It seeks to structure, classify, and compare the various approaches and techniques employed in past and current research. The main objective is to provide insights into the predictive power of online sentiments and the potential for improvement by using fine-grained sentiment analysis. Key themes explored in the thesis include: * **User-generated content (UGC) and its various forms:** This includes social media platforms like blogs, forums, and social networking services, as well as specific types of UGC such as online reviews and tweets. * **Online word-of-mouth (OWOM):** This refers to the spread of information and opinions about products, services, or events through online channels. * **Sentiment detection and classification:** The thesis examines different techniques for identifying and classifying sentiments expressed in online text, including machine learning algorithms and linguistic approaches. * **The predictive power of online sentiments:** This explores the extent to which online sentiments can be used to predict real-world events like stock market trends, sales volume, and box office revenues. * **Fine-grained sentiment analysis:** The thesis investigates the potential benefits of using more nuanced sentiment classification techniques that go beyond simple positive/negative polarity and consider specific emotions and other subtle aspects of sentiment.

Chapter Summaries

This section will focus on providing summaries of the main chapters, excluding the conclusion and any sections containing major revelations or spoilers. The summaries will focus on the key themes, arguments, or narrative elements of each chapter. * **Chapter 1: Introduction:** This chapter introduces the concept of user-generated content (UGC) and its growing influence on various aspects of society, particularly in the realm of business and marketing. It highlights the importance of understanding and analyzing sentiments expressed online and the potential for leveraging these sentiments to predict real-world outcomes. The chapter also outlines the specific research questions that will be addressed in the thesis. * **Chapter 2: Structure of Thesis:** This chapter provides a detailed overview of the thesis structure, outlining the main chapters and how they contribute to addressing the research questions. It also presents a research framework that will guide the analysis and discussion throughout the thesis. * **Chapter 3: The Need of Automated Prediction Using Online Sentiments:** This chapter explores the rationale behind using automated prediction methods based on online sentiments. It discusses the limitations of human-based predictions, highlighting biases and inefficiencies. The chapter also presents the advantages of automated prediction methods, emphasizing their objectivity, cost-efficiency, and potential to outperform human predictions. * **Chapter 4: What are the Different Prediction and Sentiment Detection Approaches and Techniques based on User-Generated-Content?:** This chapter delves into the technical background of online sentiments and the various approaches for detecting and classifying them. It first explores different types of user-generated content platforms, including social media, online communities, social networking services, weblogs, and review sites. The chapter then introduces the concept of online word-of-mouth (OWOM) and its different forms. Finally, it provides a detailed discussion of different sentiment classification techniques, including machine learning algorithms like Naïve Bayes, Maximum Entropy, and Support Vector Machines, as well as linguistic approaches like Pointwise Mutual Information and Information Retrieval (PMI-IR) and Latent Semantic Analysis (LSA). * **Chapter 5: How Consistent are Prediction Results Based on Online Sentiments?:** This chapter examines the predictive power of online sentiments by analyzing existing research on stock market prediction, sales volume prediction, and box office revenue prediction. The chapter explores the specific data sources, prediction methods, and findings of various studies, and discusses their consistency and limitations. It also identifies key challenges in building successful prediction models based on online sentiments, including the selection of appropriate sentiment metrics, data sources, and classification tools. * **Chapter 6: Do Fine-Grained Sentiments Generate New Insights and Better Prediction Results Than Coarse Sentiments?:** This chapter explores the potential benefits of using fine-grained sentiment analysis techniques, which go beyond simple positive/negative polarity and consider specific emotions and other subtle aspects of sentiment. The chapter analyzes existing research on fine-grained sentiment analysis and compares its performance with coarse sentiment approaches. It also discusses the potential for fine-grained sentiment analysis to generate new insights and improve prediction accuracy.

Keywords

The key terms and focus topics of this thesis are: online sentiments, user-generated content, social media, online word-of-mouth, sentiment detection, sentiment classification, prediction, stock market prediction, sales volume prediction, box office revenue prediction, fine-grained sentiment analysis, coarse sentiment analysis, machine learning, linguistic approaches. These keywords encompass the primary themes and concepts of the work, covering research focuses, important themes, and core concepts.

Frequently Asked Questions

Can online sentiments be used to predict real-world events?

Yes, the thesis analyzes how user-generated content and sentiments can be used to predict outcomes like stock market trends, sales volumes, and box office revenues.

What is the difference between valence and emotions in sentiment analysis?

Valence typically refers to simple positive or negative polarity, while emotions involve fine-grained analysis of specific human feelings expressed in text.

Which machine learning techniques are discussed for sentiment classification?

The research covers several techniques including Naïve Bayes, Maximum Entropy, and Support Vector Machines (SVM).

How does fine-grained sentiment analysis improve prediction?

Fine-grained sentiments go beyond simple polarity to capture nuance, which research suggests can significantly improve the accuracy of automated prediction models.

What are the limitations of using online sentiments for prediction?

Limitations include the consistency of results across different data sources, the choice of classification tools, and the scarcity of empirical data for complex emotional approaches.