Books High-performance Algorithmic Trading using Machine Learning: Building automated trading strategies with AutoML and feature engineering by Franck Bardol

"High-performance Algorithmic Trading using Machine Learning: Building Automated Trading Strategies with AutoML and Feature Engineering" by Franck Bardol is a practical guide focused on developing production-grade, machine learning-based trading systems from scratch, targeting advanced independent practitioners, quants, and developers. The book goes beyond simplistic approaches like basic return sign classification, emphasizing field-tested techniques and real-world trading scenarios.

Key aspects covered include:

Foundations of Algorithmic and Systematic Trading: Clear introductions to industry players, strategy types (momentum, statistical arbitrage, HFT), and how machine learning fits into modern trading pipelines.

Practical ML Techniques: Step-by-step pipelines for supervised learning (for advanced feature engineering and model building) and unsupervised learning (for pattern mining and anomaly detection).

Feature Engineering: Extensive coverage of designing robust features from raw data, including using accounting data, constructing custom bars (volume, volatility), and filtering out market noise.

AutoML and Low-Code Tools: Guidance on automating feature selection, model tuning, and deployment using tools like H2O and Microsoft FLAML.

Handling Unstructured Data: Techniques for extracting trading signals from news, financial reports, and alternative data using NLP.

Model Lifecycle and Backtesting: End-to-end system design covering data ingestion, backtesting, forward testing, and robust metrics/reporting, with Python code ready for real-world use.

Pattern Mining and Signal Processing: Designing ultra-fast pattern-matching algorithms and incorporating rule-based trading strategies.

Portfolio Construction: Application of advanced ML for constructing and optimizing trading portfolios.

Chapters progress from the basics of algorithmic trading through advanced supervised/unsupervised ML, ending with portfolio construction and management. The book is entirely practical, focusing on ready-to-use code, pipelines, and tools instead of heavy mathematical theory. By its conclusion, readers are equipped to design, test, and deploy institution-grade algorithmic trading strategies using state-of-the-art ML, AutoML, and feature engineering techniques.