Machine Learning in Financial Markets: Dimension Reduction and Support Vector Machine

Abstract

The chapter focuses mainly on two aspects of machine learning. The first aspect is based on classification methods such as support vector machine (SVM) and its applications in algorithmic trading based on statistical arbitrage. The second aspect is based on dimension reduction, which is the core of many solutions for problems in modern finance since big data creates its unique challenges. To achieve this goal in this chapter, different paradigms are reviewed, and logical connections between them are mentioned. It is shown how dimension reduction techniques, as well as classification methods such as SVM, is the basis for problems such as portfolio selection and optimization, statistical arbitrage, scenario generation and algorithmic trading.KeywordsDimension reductionMachine learningFinancial marketsBig dataKernel PCASupport vector machineLocal linear mappingMDSIsomapLaplacian eigenmapsPortfolio selectionScenario generationAlgorithmic tradingSVMSIRLasso