All episodes of Chat With Traders x Quantopian mini-series are available here.
In practice, no one trading model will ever be that good on its own. Luckily statistics has come up with a lot of theory about how you can combine weaker models to create better overall predictions. We’ll discuss how to combine many different trading signals into overall models and some of the practical considerations in doing so.
Sponsored by DataCamp:
- For new and intermediate programmers, DataCamp.com have an evergrowing library of interactive courses that’ll help you to become a data science wizard.
Topics of discussion:
- Why you shouldn’t expect to make money on a single alpha factor.
- Various techniques used to aggregate multiple alpha factors in one portfolio.
- How many alpha factors is too many—at what point do returns begin diminishing?
Links and resources mentioned:
- Multiple Linear Regression (Quantopian lecture)
- Position Concentration Risk (Quantopian lecture)
- Adaboost (Quantopian forum)
How to support this podcast:
- For a quick and easy way to support this free podcast, please write an honest review in iTunes. It’ll take you two minutes, and it helps massively—thank you.