Machine Learning for Finance

Kalman filters

Kalman filters are a method of extracting a signal from either noisy or incomplete measurements. They were invented by Hungarian-born, American engineer, Rudolf Emil Kalman, for the purpose of electrical engineering, and were first used in the Apollo Space program in the 1960s.

The basic idea behind the Kalman filter is that there is some hidden state of a system that we cannot observe directly but for which we can obtain noisy measurements. Imagine you want to measure the temperature inside a rocket engine. You cannot put a measurement device directly into the engine, because it's too hot, but you can have a device on the outside of the engine.

Naturally, this measurement is not going to be perfect, as there are a lot of external ...