Closed ghost closed 2 years ago
It was proper as initial_state_mean taken as 0, if you want to start your Kalman Filter guess with your initial measurement then your can change initial_state_mean (in this example approximately 1410)
Thanks, closing for now, please feel free to reopen if you have further questions.
The example in denoising data with Kalman Filter seems that is not applied properly. The initial value is odd (relates to the guess? )
https://github.com/stefan-jansen/machine-learning-for-trading/blob/main/04_alpha_factor_research/03_kalman_filter_and_wavelets.ipynb
Here is the pic from the book (second version, Kindle edition)