Applications of Computational Intelligence in Data-Driven Trading

ISBN-13: 9781119550501 / Angielski / Twarda / 2019 / 304 str.

Cris Doloc

Spis treści

About the Author xviiAcknowledgments xixAbout the Website xxiIntroduction xxiiiMotivation xxivTarget Audience xxviBook Structure xxvii1 The Evolution of Trading Paradigms 11.1 Infrastructure-Related Paradigms in Trading 11.1.1 Open Outcry Trading 21.1.2 Advances in Communication Technology 21.1.3 The Digital Revolution in the Financial Markets 31.1.4 The High-Frequency Trading Paradigm 51.1.5 Blockchain and the Decentralization of Markets 61.2 Decision-Making Paradigms in Trading 71.2.1 Discretionary Trading 81.2.2 Systematic Trading 81.2.3 Algorithmic Trading 91.3 The New Paradigm of Data-Driven Trading 11References 142 The Role of Data in Trading and Investing 152.1 The Data-Driven Decision-Making Paradigm 152.2 The Data Economy is Fueling the Future 172.2.1 The Value of Data - Data as an Asset 182.3 Defining Data and Its Utility 202.4 The Journey from Data to Intelligence 242.5 The Utility of Data in Trading and Investing 302.6 The Alternative Data and Its Use in Trading and Investing 34References 363 Artificial Intelligence - Between Myth and Reality 393.1 Introduction 393.2 The Evolution of AI 413.2.1 Early History 413.2.2 The Modern AI Era 433.2.3 Important Milestones in the Development of AI 443.2.4 Projections for the Immediate Future 483.2.5 Meta-Learning - An Exciting New Development 493.3 The Meaning of AI - A Critical View 513.4 On the Applicability of AI to Finance 543.4.1 Data Stationarity 573.4.2 Data Quality 583.4.3 Data Dimensionality 593.5 Perspectives and Future Directions 60References 624 Computational Intelligence - A Principled Approach for the Era of Data Exploration 634.1 Introduction to Computational Intelligence 634.1.1 Defining Intelligence 634.1.2 What is Computational Intelligence? 644.1.3 Mapping the Field of Study 664.1.4 Problems vs. Tools 684.1.5 Current Challenges 694.1.6 The Future of Computational Intelligence 704.1.7 Examples in Finance 714.2 The PAC Theory 724.2.1 The Probably Approximately Correct Framework 734.2.2 Why AI is a Very Lofty Goal to Achieve 754.2.3 Examples of Ecorithms in Finance 784.3 Technology Drivers Behind the ML Surge 814.3.1 Data 824.3.2 Algorithms 824.3.3 Hardware Accelerators 82References 845 How to Apply the Principles of Computational Intelligence in Quantitative Finance 875.1 The Viability of Computational Intelligence 875.2 On the Applicability of CI to Quantitative Finance 915.3 A Brief Introduction to Reinforcement Learning 945.3.1 Defining the Agent 965.3.2 Model-Based Markov Decision Process 985.3.3 Model-Free Reinforcement Learning 1015.4 Conclusions 104References 1046 Case Study 1: Optimizing Trade Execution 1076.1 Introduction to the Problem 1076.1.1 On Limit Orders and Market Microstructure 1096.1.2 Formulation of Base-Line Strategies 1116.1.3 A Reinforcement Learning Formulation for the Optimized Execution Problem 1126.2 Current State-of-the-Art in Optimized Trade Execution 1146.3 Implementation Methodology 1166.3.1 Simulating the Interaction with the Market Microstructure 1166.3.2 Using Dynamic Programming to Optimize Trade Execution 1186.3.3 Using Reinforcement Learning to Optimize Trade Execution 1196.4 Empirical Results 1226.4.1 Application to Equities 1226.4.2 Using Private Variables Only 1236.4.3 Using Both Private and Market Variables 1236.4.4 Application to Futures 1246.4.5 Another Example 1266.5 Conclusions and Future Directions 1276.5.1 Further Research 127References 1287 Case Study 2: The Dynamics of the Limit Order Book 1297.1 Introduction to the Problem 1297.1.1 The New Era of Prediction 1307.1.2 New Challenges 1317.1.3 High-Frequency Data 1327.2 Current State-of-the-Art in the Prediction of Directional Price Movement in the LOB 1337.2.1 The Contrarians 1367.3 Using Support Vector Machines and Random Forest Classifiers for Directional Price Forecast 1387.3.1 Empirical Results 1397.4 Studying the Dynamics of the LOB with Reinforcement Learning 1417.4.1 Empirical Results 1427.4.2 Conclusions 1447.5 Studying the Dynamics of the LOB with Deep Neural Networks 1457.5.1 Results 1487.6 Studying the Dynamics of the Limit Order Book with Long Short-Term Memory Networks 1497.6.1 Empirical Results 1527.6.2 Conclusions 1537.7 Studying the Dynamics of the LOB with Convolutional Neural Networks 1537.7.1 Empirical Results 1557.7.2 Conclusions 156References 1578 Case Study 3: Applying Machine Learning to Portfolio Management 1598.1 Introduction to the Problem 1598.1.1 The Problem of Portfolio Diversification 1608.2 Current State-of-the-Art in Portfolio Modeling 1618.2.1 The Classic Approach 1618.2.2 The ML Approach 1628.3 A Deep Portfolio Approach to Portfolio Optimization 1638.3.1 Autoencoders 1648.3.2 Methodology - The Four-Step Algorithm 1668.3.3 Results 1678.4 A Q-Learning Approach to the Problem of Portfolio Optimization 1678.4.1 Problem Statement 1688.4.2 Methodology 1698.4.3 The Deep Q-Learning Algorithm 1698.4.4 Results 1708.5 A Deep Reinforcement Learning Approach to Portfolio Management 1708.5.1 Methodology 1708.5.2 Data 1718.5.3 The RL Setting: Agent, Environment, andPolicy 1728.5.4 The CNN Implementation 1728.5.5 The RNN and LSTM Implementations 1728.5.6 Results 173References 1749 Case Study 4: Applying Machine Learning to Market Making 1759.1 Introduction to the Problem 1759.2 Current State-of-the-Art in Market Making 1779.3 Applications of Temporal-Difference RL in Market Making 1809.3.1 Methodology 1809.3.2 The Simulator 1819.3.3 Market Making Agent Specification 1829.3.4 Empirical Results 1859.4 Market Making in High-Frequency Trading Using RL 1899.4.1 Methodology 1909.4.2 Experimental Setting 1919.4.3 Results and Conclusions 1929.5 Other Research Studies 192References 19310 Case Study 5: Applications of Machine Learning to Derivatives Valuation 19710.1 Introduction to the Problem 19710.1.1 Problem Statement and Research Questions 19910.2 Current State-of-the-Art in Derivatives Valuation by Applying ML 20010.2.1 The Beginnings: 1992-2004 20110.2.2 The Last Decade 20210.3 Using Deep Learning for Valuation of Derivatives 20410.3.1 Implementation Methodology 20510.3.2 Empirical Results 20710.3.3 Conclusions and Future Directions 20810.3.4 Other Research Studies 20810.4 Using RL for Valuation of Derivatives 21010.4.1 Using a Simple Markov Decision Process 21010.4.2 The Q-Learning Black-Scholes Model (QLBS) 212References 21411 Case Study 6: Using Machine Learning for Risk Management and Compliance 21711.1 Introduction to the Problem 21711.1.1 Challenges 21811.1.2 The Problem 21911.2 Current State-of-the-Art for Applications of ML to Risk Management and Compliance 21911.2.1 Credit Risk 21911.2.2 Market Risk 22011.2.3 Operational Risk 22111.2.4 Regulatory Compliance Risk and RegTech 22211.2.5 Current Challenges and Future Directions 22311.3 Machine Learning in Credit Risk Modeling 22411.3.1 Data 22511.3.2 Models 22511.3.3 Results 22611.4 Using Deep Learning for Credit Scoring 22711.4.1 Introduction 22711.4.2 Deep Belief Networks and Restricted Boltzmann Machines 22811.4.3 Empirical Results 23011.5 Using ML in Operational Risk and Market Surveillance 23011.5.1 Introduction 23011.5.2 An ML Approach to Market Surveillance 23211.5.3 Conclusions 233References 23312 Conclusions and Future Directions 23712.1 Concluding Remarks 23712.2 The Paradigm Shift 23912.2.1 Mathematical Models vs. Data Inference 24012.3 De-Noising the AI Hype 24312.3.1 Why Intellectual Honesty Should Not Be Abandoned 24412.4 An Emerging Engineering Discipline 24512.4.1 The Problem 24612.4.2 The Market 24612.4.3 A Possible Solution 24612.5 Future Directions 247References 248Index 249

O autorze

CRIS DOLOC is a leading computational scientist with more than 25 years of experience in quantitative finance. He holds a PhD in Computational Physics and is currently teaching at the University of Chicago in the Financial Mathematics program. Cris is also the founder of FintelligeX, a technology platform designed to promote data-driven education, and he is very passionate about the opportunities that recent developments in Cognitive Computing and Computational Intelligence could bring to the field of Quant education.