In all books by Dr Chan, the illustrated approaches are underlined by equity curved. Which I personally really appreciate!
So before enrolling to the festure-engineering course, it would be great to see how one can (if no fee, plus some luck maybe) improve or develop a strategy?
If the feature engineering course for multi-asset (portfolio), or stst-arb or singe asset?
Thanks a lot!
The below is one of the example from the course where using data with surviorship bias caused inflated results. Similarly, there are other examples in the course which talk about the importance of good quality data and how it can improve your trading performance.
The concepts in data engineering and feature engineering course can be applied to a single asset as well as to multi-asset portfolio.
The concepts covered in this course are indifferent to the machine learning algorithms you use. But the focus is to find solutions to many shared generic problems such as data labeling and feature selection. If these things are not handled well or not paid enough attention to then irrespective of how good your machine learning algorithm is, it won't work as expected in live trading. The focus of this course is to develop framework to create robust strategies.