Hands-On Machine Learning for Algorithmic Trading

Key takeaways and lessons learned

Important insights to keep in mind as you proceed to the practice of ML for trading include:

• Data is the single most important ingredient
• Domain expertise helps realize the potential value in the data, especially in finance
• ML offers tools for many use cases that should be further developed and combined to create solutions for new problems using data
• The choice of model objectives and performance diagnostics are key to productive iterations towards an optimal system
• Backtest overfitting is a huge challenge that requires significant attention
• Transparency around black-box models can help build confidence and facilitate adoption

We will elaborate a bit more on each of these ideas.