Kalman Filter to solve time-dependent problem in time-series

Hello everyone,

I try to fit the time-series using regression with additional seasonal parameter estimation. However, the seasonal signal inconsistent and make the seasonal that used to be a signal become residual. Did anyone have solve this problem before ? and how is step to solve that ? Thank you

Answer 1

[Vidip Jain]

I understand that the seasonal signal in your time series is inconsistent and is making the seasonal that used to be a signal become residual.

There are a few possible reasons why the seasonal signal in your time series data may be inconsistent. One possibility is that the seasonal signal is non-stationary, meaning that it changes over time. Another possibility is that there are outliers in the data that are affecting the seasonal signal.

If the seasonal signal is non-stationary, then you may need to use a more sophisticated regression model that can account for non-stationarity. For example, you could use a time-varying coefficient regression model or a state space model.

If there are outliers in the data, then you may need to remove them before fitting the regression model. You can use a variety of methods to remove outliers, such as the interquartile range (IQR) method.

For further information, refer to the documentation links below:

• https://www.mathworks.com/help/matlab/ref/iqr.html
• https://www.mathworks.com/help/control/ref/ss.html