Implementing Machine Learning for Finance

Bring together machine learning (ML) and deep learning (DL) in financial trading, with an emphasis on investment management. This book explains systematic approaches to investment portfolio management, risk analysis, and performance analysis, including predictive analytics using data science procedures.

The book introduces pattern recognition and future price forecasting that exerts effects on time series analysis models, such as the Autoregressive Integrated Moving Average (ARIMA) model, Seasonal ARIMA (SARIMA) model, and Additive model, and it covers the Least Squares model and the Long Short-Term Memory (LSTM) model. It presents hidden pattern recognition and market regime prediction applying the Gaussian Hidden Markov Model. The book covers the practical application of the K-Means model in stock clustering. It establishes the practical application of the Variance-Covariance method and Simulation method (using Monte Carlo Simulation) for value at risk estimation. It also includes market direction classification using both the Logistic classifier and the Multilayer Perceptron classifier. Finally, the book presents performance and risk analysis for investment portfolios.

By the end of this book, you should be able to explain how algorithmic trading works and its practical application in the real world, and know how to apply supervised and unsupervised ML and DL models to bolster investment decision making and implement and optimize investment strategies and systems.

WHAT YOU WILL LEARN

* Understand the fundamentals of the financial market and algorithmic trading, as well as supervised and unsupervised learning models that are appropriate for systematic investment portfolio management
* Know the concepts of feature engineering, data visualization, and hyperparameter optimization
* Design, build, and test supervised and unsupervised ML and DL models
* Discover seasonality, trends, and market regimes, simulating a change in the market and investment strategy problems and predicting market direction and prices
* Structure and optimize an investment portfolio with preeminent asset classes and measure the underlying risk

WHO THIS BOOK IS FOR

Beginning and intermediate data scientists, machine learning engineers, business executives, and finance professionals (such as investment analysts and traders)TSHEPO CHRIS NOKERI harnesses big data, advanced analytics, and artificial intelligence to foster innovation and optimize business performance. In his functional work, he has delivered complex solutions to companies in the mining, petroleum, and manufacturing industries. He initially completed a bachelor’s degree in information management. He then graduated with an honors degree in business science at the University of the Witwatersrand on a TATA Prestigious Scholarship and a Wits Postgraduate Merit Award. They unanimously awarded him the Oxford University Press Prize. He has authored the Apress book Data Science Revealed: With Feature Engineering, Data Visualization, Pipeline Development, and Hyperparameter Tuning.

Chapter 1: Introduction to the Financial Markets and Algorithmic Trading

Foreign exchange market

- Exchange rate

- Exchange rates quotation

The Interbank market

The retail market

Brokerage

- Understanding leverage and margin

- Contract for difference trading

The share market

Raising capital

- Public listing

- Stock exchange

- Share trading

Speculative nature of foreign exchange market

Techniques for speculating market movement

Algorithmic trading

- Supervised machine learning

The parametric method

- The non-parametric method

Binary classification

Multiclass classification

- The ensemble method

- Unsupervised learning - Deep learning

- Dimension reduction

Chapter 2: Forecasting Using ARIMA, SARIMA and Additive Model

Time series in action

Split data into training and test data

Test for stationary

Test for white noise

Autocorrelation function

Partial autocorrelation function

The moving averages smoothing technique

The exponential smoothing technique

Rate of return

The ARIMA Model

ARIMA Hyperparameter Optimization

- Develop the ARIMA model

- Forecast prices using the ARIMA model

The SARIMA model

- Develop SARIMA model

- Forecast using the SARIMA model

Additive model

- Develop the additive model

- Forecast prices the additive model

- Seasonal decomposition

Conclusion

Chapter 3: Univariate Time Series using Recurrent Neural Nets

What is deep learning?

Activation function

Loss function

Optimize an artificial neural network

The sequential data problem

The recurrent net model

The recurrent net problem

The LSTM model

Gates

Unfolded LSTM network

Stacked LSTM network

LSTM in action

- Split data into training, test and validation

- Normalize data

- Develop LSTM model

- Forecasting using the LSTM

- Model evaluation

- Training and validation loss across epochs - Training and validation accuracy across epochs

Conclusion

Chapter 4: Discover Market Regimes

HMM

HMM application in finance

- Develop GaussianHMM

Mean and variance

Expected returns and volumes

Conclusions

Chapter 5: Stock Clustering

Investment Portfolio Diversification

Stock market volatility

K-Means clustering

K-Means in practice

Conclusions

Chapter 6: Future Price Prediction using Linear Regression

Linear Regression in Practice

Detect missing values

Pearson correlation

Covariance

Pairwise scatter plot

Eigen matrix

Split data into training and test data.

Normalize data

Least squares model hyperparameter optimization

Step 1: Fit least squares model with default hyperparameters

Step 2: Determine the mean and standard deviation of the cross-validation scores

Step 3: Determine Hyper-parameters that yield the best score.

Develop least squares model

Find an intercept

Find the estimated coefficient

Test least squares model performance using SciKit-Learn

Plotting actual values and predicted values

Conclusion

Chapter 7: Stock Market Simulation

Understanding value at risk

Estimate VAR using the Variance-Covariance Method

Understanding Monte Carlo

Application of Monte Carlo simulation in finance

- Run Monte Carlo simulation

- Plot simulations

Conclusions

Chapter 8: Market Trend Classification using ML and DL

Classification in practice

Data preprocessing

Split Data into training and test data

Logistic regression

- Finalize a logistic classifier

- Evaluate a logistic classifier

- Learning curve

Multilayer layer perceptron

- Architecture

- Finalize model

- Training and validation loss across epochs

- Training and validation accuracy across epochs

Conclusions

Chapter 9: Investment Portfolio and Risk Analysis

Investment

Investment Analysis

Investment Risk Management

Investment Portfolio Management Pyfolio in action

Performance statistics

Drawback

Rate of returns

Annual rate of return

Rolling returns

- Monthly rate of returns

Conclusions