Machine Learning for the Markets

This practical two-day workshop shows how machine learning techniques can be used to identify predictive variables and trading rules. Delegates will learn the main machine learning algorithms and make extensive use of MATLAB’s Statistics and Machine Learning Toolbox.

Each delegate will be equipped with a PC to perform the machine learning exercises. No prior MATLAB experience is required.

11 - 12 May 2020

Duration:

2 Days

Location:

London, UK – Tower Hotel, London E1

Trainer:

Ernest Chan

Course Fee:

£1990 + VAT

Register

OVERVIEW OF MACHINE LEARNING TECHNIQUES

The machine learning paradigm

  • Features selection
  • Training vs test sets
  • Cross validation
  • Boot-strapping
  • Data snooping bias

Unsupervised Learning/Supervised Learning

  • Overview of main classification algorithms
  • Structured vs Unstructured Data
  • Issues with both types of data
  • Setting up the problem with multiple linear regression as the learning model
  • Exercise: predict 1-day SPY return using simple technical indicators

LEARNING ALGORITHMS

  • Stepwise linear regression
  • Classification and regression trees (CART)
  • Stopping criteria for tree growing
  • Using the whole tree or selecting certain nodes for prediction?
  • Reducing overfitting by cross-validation
  • Increasing training sample size by bootstrapping/bagging
  • Learning from past errors: boosting
  • Which technique gives the most accurate predictions?

Support vector machine (SVM)

  • Predicting the sign of returns

Neural networks (NN)

  • Neural network as nonlinear function fitting
  • What network architecture to pick?
  • Drawbacks of using NN for financial predictions
  • Avoiding overfitting: dropout technique

Extended exercise: predicting SPY returns using various learning algorithms

EXTENDED EXERCISE ON FEATURES SELECTION

  • Building a multifactor stock selection model using fundamental factors
  • Techniques: multiple regression, stepwise regression, and CART
  • What fundamental factors are most useful for predicting stock portfolio returns?