From Valence to Emotions

Table of Contents

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

Objective & Topics

This thesis investigates the potential of predicting real-world outcomes using user-generated content (UGC), specifically examining how online sentiments—expressed as valences or emotions—can serve as predictive indicators. The central research objective is to structure, classify, and compare existing sentiment-based prediction literature to determine the consistency of research findings and the added value of fine-grained sentiment analysis over traditional coarse classification.

• Theoretical framework of UGC platforms and sentiment detection techniques.
• Classification of empirical research into three specific prediction domains: stock markets, sales volume, and box office revenues.
• Critical analysis of prediction methods, consistency, and limitations within current empirical literature.
• Comparison of coarse sentiment classification (positive/negative valence) versus fine-grained approaches (emotion tracking).
• Evaluation of predictive accuracy in the context of different data sources and analytical models.

Excerpt from the Book

Summary of Chapters

1 Introduction: Provides an overview of the growth of user-generated content and outlines the thesis's objective to analyze and compare sentiment-based prediction methods.

2 Structure of Thesis: Outlines the research approach, including the theoretical background and the subsequent classification of empirical literature into specific prediction subjects.

3 The Need of Automated Prediction Using Online Sentiments: Discusses the advantages of IT-based sentiment analysis over human judgment, focusing on cost-efficiency and objectivity.

4 What are the Different Prediction and Sentiment Detection Approaches and Techniques based on User-Generated-Content?: Explains the technical background of UGC, online word-of-mouth (OWOM), sentiment forms, and computational classification tools.

5 How Consistent are Prediction Results Based on Online Sentiments?: Analyzes and compares empirical studies on predicting stock markets, sales volume, and box office revenues, highlighting method consistency and limitations.

6 Do Fine-Grained Sentiments Generate New Insights and Better Prediction Results Than Coarse Sentiments?: Investigates whether advanced emotion-based sentiment analysis leads to improved predictive accuracy compared to simple valence classification.

7 Conclusion: Summarizes the findings regarding the research questions and identifies core challenges in sentiment-based prediction modeling.

8 Managerial Implications: Offers insights for researchers and practitioners on leveraging sentiment-based predictions for business applications like stock planning and product development.

Keywords

User-Generated Content, UGC, Online Word-of-Mouth, OWOM, Sentiment Analysis, Sentiment Classification, Machine Learning, Naïve Bayes, Support Vector Machines, Latent Semantic Analysis, Predictive Analytics, Stock Market Prediction, Sales Volume, Box Office Revenue, Fine-grained Sentiment

Frequently Asked Questions

What is the core focus of this thesis?

The thesis focuses on how online sentiments found in user-generated content (UGC) can be used to predict real-world outcomes across different industries and topics.

What are the primary subject areas analyzed?

The research primarily evaluates prediction models for stock market developments, product sales volume, and box office revenues.

What is the main goal of this research?

The primary goal is to structure, classify, and compare existing literature on sentiment-based prediction to understand its consistency and assess if fine-grained sentiment analysis provides better results than coarse methods.

What scientific methodologies are utilized in the reviewed studies?

The studies reviewed utilize various methods, including machine learning techniques (e.g., Naïve Bayes, Support Vector Machines, Maximum Entropy) and linguistic approaches (e.g., Latent Semantic Analysis, Pointwise Mutual Information).

What is addressed in the main body of the work?

The main body establishes the theoretical framework of UGC, defines different forms of sentiment (volume, valence, emotions), and provides a comprehensive, domain-specific evaluation of empirical studies.

Which terms best characterize this work?

Key terms include User-Generated Content (UGC), Online Word-of-Mouth (OWOM), sentiment classification, predictive analytics, and fine-grained versus coarse sentiment metrics.

How do fine-grained sentiments differ from coarse sentiments?

Coarse sentiments generally involve binary classification (positive or negative), while fine-grained sentiments capture more complex emotional dimensions like anxiety, joy, or specific mood states, which offer deeper insights.

Does the thesis conclude that automated sentiment prediction is effective?

Yes, the thesis finds that in the majority of analyzed studies, sentiment data holds predictive value, though results vary depending on the chosen data source and the specific prediction subject.