Content: I: INTRODUCTIONIntroductionBackgroundWhat Is Online Portfolio Selection?MethodologyBook OverviewProblem FormulationProblem Settings Transaction Costs and Margin Buying ModelsEvaluationSummary II: Principles BenchmarksBuy-and-Hold StrategyBest Stock Strategy Constant Rebalanced PortfoliosFollow the Winner Universal PortfoliosExponential Gradient Follow the Leader Follow the Regularized LeaderSummary Follow the LoserMean ReversionAnticorrelation Summary Pattern Matching Sample Selection Techniques Portfolio Optimization Techniques Combinations Summary Meta-Learning Aggregating AlgorithmsFast Universalization Online Gradient and Newton Updates Follow the Leading History Summary III: Algorithms Correlation-Driven Nonparametric Learning PreliminariesFormulations AlgorithmsAnalysisSummary Passive-Aggressive Mean Reversion Preliminaries FormulationsAlgorithms AnalysisSummaryConfidence-Weighted Mean ReversionPreliminaries FormulationsAlgorithms AnalysisSummary Online Moving Average Reversion PreliminariesFormulations Algorithms Analysis Summary IV: Empirical Studies Implementations The OLPS PlatformData SetupsPerformance Metrics SummaryEmpirical Results Experiment 1: Evaluation of Cumulative Wealth Experiment 2: Evaluation of Risk and Risk-Adjusted Return Experiment 3: Evaluation of Parameter SensitivityExperiment 4: Evaluation of Practical Issues Experiment 5: Evaluation of Computational TimeExperiment 6: Descriptive Analysis of Assets and Portfolios SummaryThreats to Validity On Model Assumptions On Mean Reversion Assumptions On Theoretical Analysis On Back-Tests SummaryV: Conclusion Conclusions Future DirectionsAppendix A: OLPS: A Toolbox for Online Portfolio Selection IntroductionFramework and InterfacesStrategiesSummary Appendix B: Proofs and Derivations Proof of CORNDerivations of PAMRDerivations of CWMRDerivation of OLMARAppendix C: Supplementary Data and Portfolio Statistics BibliographyIndex