In this video, we are applying the third order polynomial DLM and moving average (MA) DLM in the time series of 10-year Treasury. DLM (Dynamic Linear Model) is a model based on Bayesian statistics. The fitting process could also be called as Kalman filter.

00:00 Intro
00:19 The definition of the third order polynomial DLM
08:36 The definition of ARMA component in DLM
15:32 Visualizing the 10-year US Treasury yield
18:27 Applying polynomial model
30:09 Applying polynomial and MA model
35:16 Applying polynomial, MA, and seasonal model
38:36 Model comparison

In 26:28, I made a mistake in interpreting the ACF plot, but I corrected it in the subtitle. Please take a look at the subtitle to see the right interpretation.

The code and dataset are in my Github: https://github.com/stephanielees/Bond...

#arma #movingaverage #timeseries #bayesian #datascience #treasurybonds