Alternative Data and Artificial Intelligence Techniques: Applications in Investment and Risk Management

Description

This book introduces a state-of-art approach in evaluating portfolio management and risk based on artificial intelligence and alternative data. The book covers a textual analysis of news and social media, information extraction from GPS and IoTs data, and risk predictions based on small transaction data, etc. The book summarizes and introduces the advancement in each area and highlights the machine learning and deep learning techniques utilized to achieve the goals. As a complement, it also illustrates examples on how to leverage the python package to visualize and analyze the alternative datasets, and will be of interest to academics, researchers, and students of risk evaluation, risk management, data, AI, and financial innovation. Preface Contents About the Authors List of Figures List of Tables Part I Portfolio and Risk Management Overview 1 An Introduction to Quantitative Portfolio Management and Risk Management 1.1 Introduction 1.2 Types of Portfolio Management 1.3 The Classic Asset and Derivatives 1.3.1 Classic Assets Classes in Portfolio Management 1.3.2 Derivatives in Portfolio Management 1.4 Traditional and Modern Approaches 1.4.1 Traditional Approaches to Portfolio Management 1.4.2 Modern Approaches to Portfolio Management 1.5 Tools for Measuring Portfolio Returns 1.6 Variance on Return in a Portfolio 1.7 Conclusions References 2 The Major Trends in Global Financial Asset Management 2.1 Introduction 2.2 Global Asset Management Today 2.2.1 The United States 2.2.2 Europe 2.2.3 China 2.3 Development Trends in the Asset Management Industry 2.3.1 ESG Investing 2.3.2 Blockchain 2.3.3 Robo-Advisors References Part II Machine Learning and Alternative Data Overview 3 Machine Learning and AI in Financial Portfolio Management 3.1 Overview 3.1.1 Basic Introduction to Machine Learning 3.1.2 Overview of the Application of Machine Learning in Financial Portfolio Management 3.1.2.1 Mobile Payments 3.1.2.2 P2P 3.1.2.3 Big Data Analysis 3.1.2.4 Digital Currency and Data Blockchain Technology 3.1.2.5 Intelligent Trading and Finance 3.1.3 Implementation Conditions for Machine Learning 3.1.3.1 Big Data 3.1.3.2 Techniques 3.1.3.3 Humans 3.1.4 Introduction to Article Structure 3.2 Analysis of Machine Learning Application 3.2.1 Supervised Learning 3.2.1.1 Classification 3.2.1.2 Logistic Regression 3.2.1.3 Support Vector Machines (SVMs) 3.2.1.4 Decision Trees 3.2.1.5 Random Forest 3.2.1.6 Hidden Markov Models (HMMs) 3.2.1.7 Regression 3.2.1.8 Nonparametric Regression: Loess and K-Nearest Neighbor 3.2.1.9 Dynamical Systems: Kalman Filtering 3.2.1.10 Extreme Gradient Boosting 3.2.2 Unsupervised Learning 3.2.2.1 Clustering 3.2.2.2 K-Means 3.2.2.3 Birch 3.2.2.4 Ward’s Method 3.2.2.5 Factor Analysis Through PCA 3.2.3 Deep Learning 3.2.3.1 Multilayer Perceptron (MLP) 3.2.3.2 Time-Series Analysis: Long Short-Term Memory (LSTM) 3.2.3.3 Convolutional Neural Networks (CNNs) 3.2.3.4 Restricted Boltzmann Machines (RBM) 3.2.4 Reinforcement Learning 3.2.5 Active Learning 3.3 Comparison of Machine Learning Algorithms 3.3.1 Comparison of Supervised Learning Algorithms—Regression 3.3.2 Comparison of Supervised Learning Algorithms—Classification 3.3.3 Comparison of Unsupervised Learning Algorithms—Clustering 3.4 Select the Best Model 3.4.1 Variance-Bias Trade-Off Theory 3.4.2 Model Complexity 3.5 Application of Machine Learning in Financial Field 3.6 Problem Analysis of Machine Learning 3.7 Future Perspectives 4 Introduction of Alternative Data in Finance 4.1 Alternative Data Overview 4.1.1 Trends 4.1.2 Risk 4.2 Sources of Alternative Data 4.2.1 Data Generated by Personal Activities 4.2.2 Data Generated by Business Activities 4.2.3 Data Obtained by High-Tech Monitoring 4.3 Criteria for Evaluating Alternative Datasets 4.4 Working with Alternative Data 4.4.1 Application Example: Google 4.4.2 Application Example: Peloton 4.4.2.1 Advantage Analysis Using Alternative Data 4.4.3 Application Example: WeWork References 5 Alternative Data Utilization from a Country Perspective 5.1 The United States 5.1.1 Microsoft 5.1.2 BondCliQ 5.1.3 EventVestor 5.1.4 Verbatim 5.1.5 Thinknum 5.1.6 IBM 5.1.7 Oracle 5.1.8 Google 5.1.9 Amazon 5.1.10 HP Vertica 5.1.11 Intel 5.1.12 Teradata 5.2 China 5.2.1 CDP (Cloud Data Platform) 5.2.2 Huawei 5.2.3 Alibaba Group 5.2.4 Super Pair Technology 5.2.5 Choice 5.2.6 SmarTag 5.2.7 Jove Bird 5.2.8 IUAP Yonyoucloud 5.2.9 Sugon 5.3 Europe 5.3.1 PatentSight 5.3.2 Peekd 5.3.3 Unacast 5.3.4 Owlin 5.3.5 EPFR 5.3.6 Huq Industries (Geo-Location) 5.4 Asia (Except China) 5.4.1 Nikkei 5.4.2 Qmit 5.4.3 Datapulse 5.4.4 Liases Foras 5.4.5 Propstack References Part III Factors Applications in Financial Management 6 Smart Beta and Risk Factors Based on Textural Data and Machine Learning 6.1 Introduction 6.2 Textural Analysis Technologies 6.3 Natural Language Processing 6.4 Machine Learning/Deep Learning (ML/DL) 6.5 Factors for Finance Built on Textural Dataset Analysis 6.5.1 Readability 6.5.2 Tone and Sentiment Factors 6.5.2.1 Tone Factors of Disclosures from Listed Company 6.5.2.2 Tone Factors of Media 6.5.2.3 Sentimental Factors of Investor Community and Social Media 6.5.2.4 Sentimental Factors of Employee Ratings 6.5.2.5 Sentiment Factors of Customer Ratings 6.5.3 Similarity Factors 6.5.4 Semantic Factors 6.5.4.1 Semantic Factors of Disclosures from Listed Companies 6.5.4.2 Semantic Factors of Regulatory Comments 6.5.4.3 Semantic Factors of Alternative Data 6.5.5 Uncertainty Factors 6.5.6 Accuracy Factors 6.5.7 Popularity Factors 6.5.7.1 Popularity Factors of Firm Disclosure 6.5.7.2 Popularity Factors of Market Trends 6.6 Conclusion References 7 Smart Beta and Risk Factors Based on IoTs 7.1 Introduction 7.2 A Risk Assessment Model Based on IoT and AIoT 7.2.1 Analytic Hierarchy Process 7.2.2 Artificial Immune Models 7.2.3 The Cloud Transformation Model 7.3 Applications of IoT and AIoT in Finance 7.3.1 Malware Capture 7.3.2 In the Field of Credit 7.3.3 In the Insurance Field 7.3.4 In Operations Monitoring References 8 Environmental, Social Responsibility, and Corporate Governance (ESG) Factors of Corporations 8.1 Introduction of Environmental, Social, and Governance (ESG) 8.1.1 The Development of ESG 8.1.2 The Definition of ESG 8.2 ESG in the Eyes of Investors 8.2.1 Investors Positive About ESG Investing 8.2.2 Investors Negative About ESG Investing 8.2.3 Regional Differences in Investor Perspectives on ESG Investing 8.2.4 E, S, G in the Eyes of Investors Respectively 8.3 The Influence of ESG on Firm Risk 8.3.1 Systematic Risks 8.3.2 Idiosyncratic Risks 8.4 The Influence of ESG on Firm Performance and Firm Value 8.4.1 ESG Ratings 8.4.2 Why Does ESG Matter? Starting from the DCF Model 8.4.3 The Influence of ESG on Discounted Rate: Firm Risk 8.4.3.1 ESG Reduces the company’s Discount Rate by Reducing Systematic Risk 8.4.3.2 ESG Reduces the company’s Discount Rate by Reducing Firm Idiosyncratic Risk 8.4.4 The Influence of ESG on Discounted Rate: Cost of Capital 8.4.5 Other Channels 8.5 Is ESG a Risk Factor? 8.6 The Digital Economy and ESG References 9 Sentiment Factors in Finance 9.1 What Is Sentiment Factor? 9.1.1 An Introduction to Sentiment Analysis 9.1.2 Literature Review on Sentiment Analysis 9.1.3 An Example of Sentiment Affecting the Market 9.2 Investor Sentiment and Behavioral Finance 9.2.1 Limits to Arbitrage 9.2.2 Psychology 9.2.3 Herding Behavior 9.3 Sentiment’s Market Influence 9.3.1 Sentiment in the Stock Market 9.3.2 Sentiment in the Cryptocurrency Market 9.4 Sentiment Factor Constructions and Sentiment Analysis 9.4.1 Sentiment Metrics 9.4.2 Sentiment Analysis Methods References Part IV Case Studies of Machine Learnings and Alternative Data 10 Fraud and Deception Detection: Text-Based Data Analytics 10.1 Copycat Detection 10.1.1 Estimation Results 10.1.1.1 The Aggregate Effect of Copycats on the Demand of the Original 10.1.1.2 How the Quality and the Imitation Type of the Copycats Affect the Original App’s Demand 10.2 Fraudulent Reviews 10.2.1 How to Identify False Comments Through Text Information 10.2.2 How to Deal with Potentially False Comments 10.2.2.1 Support Vector Machine 10.2.2.2 Naive Bayes 10.2.2.3 Random Forest 10.2.2.4 Adaptive Boost References 11 Machine Learning Technique in Trading: A Case Study in the EURUSD Market 11.1 Introduction to Foreign Exchange Markets 11.2 Characteristics of Foreign Exchange Markets 11.2.1 24 Hours Trading × 5 Days a Week 11.2.2 Market Transparency 11.2.3 Highly Leveraged Market 11.2.4 Higher Liquidity 11.3 Euro Dollar Exchange Rate (EURUSD) 11.4 Fundamental Factors Affecting the Foreign Exchange Rate 11.5 Data and Trading Strategy Overview 11.5.1 Data Used for Modeling a Trading Strategy in the EURUSD Market 11.5.2 Benchmark Strategies 11.6 Supervised Machine Learning Techniques 11.6.1 Random Forest 11.6.2 Support Vector Machines 11.6.3 K-Nearest Neighbors 11.7 Trading Strategy 11.7.1 The Overall Performance of All the Techniques 11.7.1.1 Cumulative Returns for the Test Period 11.7.2 Performance During the 2008 Crisis 11.8 Conclusion References 12 Analyzing the Special Purpose Acquisition Corporation (SPAC) with ESG Factors 12.1 Brief Introduction to SPACs 12.1.1 The History and Background of SPACs 12.1.2 The Operating Mechanism of SPACs 12.1.3 The Characteristics of SPACs 12.2 The Role of SPACs 12.2.1 The Importance of Founders 12.2.2 The Development Status of SPACs 12.2.3 Regulatory Norms and Post-Merger Performance 12.2.3.1 Post-Merger Performance 12.2.3.2 Regulatory Norms 12.3 Analysis of the Impact of Founder Factors on the Revenue of SPACs 12.3.1 Data Source and Description 12.3.2 Data Processing 12.3.2.1 Technical Difficulties: Parsing HTML Files 12.3.2.2 Gain Educational Experience 12.3.3 Data Analysis 12.3.3.1 Descriptive Analysis 12.3.3.2 Correlation Analysis 12.3.4 Data Visualization 12.4 Conclusions References 13 ESG Impacts on Corporation’s Fundamental: Studies from the Healthcare Industry 13.1 Introduction 13.2 Data and Methodology 13.2.1 Financial Data 13.2.2 ESG Data 13.2.3 Transforming the ESG Data 13.2.4 Formulating the ESG Score 13.2.5 Company Selection 13.2.6 Variables and Summary Statistics 13.3 Empirical Model and Results 13.3.1 Has the Pharmaceutical Industry Gained Excess Economic Benefits from the Epidemic? 13.3.2 Do ESG Factors Affect Firms’ Economic Performance on the ROA/ROE Level? 13.4 Investment Strategy on ESG Factors 13.4.1 Do ESG Scores Generate Excess Returns? 13.4.2 Sensitivity Testing 13.5 Conclusion Appendix 1: List of Companies Used in the Research Appendix 2: Quantile ESG Score for Every Quarter Appendix 3: Results of the Return of Strategy and Control Group Appendix 4: Cumulative Return of Strategy and Control Group Appendix 5: Quarterly Return of 3 Strategies and Control Group Appendix 6: Cumulative Return of 3 Strategies and Control Group References Part V Techniques in Data Visualization and Database 14 Data Visualization 14.1 Data Visualization Fundamentals 14.1.1 Background of Data Visualization 14.1.2 Research Status 14.2 Introduction to Python Visualization Tools 14.2.1 Matplotlib 14.2.1.1 Introduction 14.2.1.2 Basic Application 14.2.2 Seaborn 14.2.2.1 Introduction 14.2.2.2 Basic Application 14.2.3 Plotly 14.2.3.1 Introduction 14.2.3.2 Basic Application 14.2.4 Pyecharts 14.2.4.1 Introduction 14.2.4.2 Basic Application 14.3 Data Distribution Chart 14.3.1 Statistical Histogram 14.3.2 Boxplot 14.3.3 Scatter Chart and Line Chart 14.3.4 Column Distribution Chart 14.3.5 Violin Chart 14.4 Financial Data Case Analysis 14.4.1 Data Sources 14.4.2 Financial Chart 14.5 Summary References 15 Interacting with a MongoDB Database from a Python Function in AWS Lambda 15.1 MongoDB 15.1.1 Introduction to MongoDB 15.1.2 MongoDB Pricing 15.1.3 Creating a MongoDB Database 15.1.3.1 Tone Factors Based on Disclosures from Listed Companies 15.2 Python 15.2.1 Prepare a Python Environment 15.2.2 Simple Example Code to Test Locally 15.2.3 Modified Code for AWS 15.2.4 Downloading Python Dependencies 15.3 AWS 15.3.1 Introduction to AWS 15.3.2 Introduction to AWS Lambda 15.3.3 AWS Pricing 15.3.4 AWS IAM Accounts vs. AWS Root Accounts 15.3.5 Creating an IAM User 15.3.6 Creating an AWS Lambda Function 15.3.7 Adding Triggers to an AWS Lambda Function References Index