Seasonality in Time Series

Seasonality refers to patterns that repeat at fixed intervals (daily, weekly, monthly, or yearly). Example: higher sales of cold drinks during summer.

Ways to Handle Seasonality

1. Seasonal Differencing Remove seasonal effects by subtracting the value from the same season in a previous cycle:

   $y_t^{\prime }=y_t-y_{t-s}$

   where $s$ is the seasonal period (e.g., $s=12$ for monthly data with yearly seasonality).

2. Break the series into trend (T), seasonal (S), and residual (R) components:

   1. Decomposition in Time Series
   2. Additive vs Multiplicative Models Time Series
   3. STL Decomposition

3. Seasonal Dummy Variables Encode seasonality as categorical indicators (e.g., month or quarter dummies) to use as features in machine learning models.