Ernest Chan, QTS Capital Management, LLC
The traditional paradigm of applying nonlinear machine learning techniques to algorithmic trading strategies typically suffers massive data snooping bias. On the other hand, linear techniques, inspired and constrained by in-depth domain knowledge, have proven to be valuable. This presentation describes the application of the Kalman filter, a quintessentially linear technique, in two different ways to algorithmic trading.
Recorded: 19 September 2013