torreblanca99 / auto_regressive_process-

Learning to identify an Auto-Regressive model (AR) of second order.

Simulating a Auto-Regressive process in time series

The purpose of this reppository is learning to identify an Auto-Regressive process. In order to do that, the definition and the process to identify this sort of series is presented. The next diagram shows the general process:

An AR(2) process will be simulated using the equation: $$y_t = 0.33 y_{t-1} + 0.50 y_{t-2} $$ The actions to execute are:

1. Gather data: Simulate de AR(2) model with the above equation.
2. Test stationary.
   • Apply transformations until our series is stationary.
3. Plot ACF: Is it slowly decaying?
4. Plot the PACF:get the order of our AR model.
5. Make forecasts over the test: use rolling forecast with a window lenght of the AR's order.
6. Plot forecasting.

Bibliography:

• Peixeiro, M. (2022). Time Series Forecasting in Python (1st ed., Chapter 4, pp. 81-100). Manning Publications.