Stationarity
Explore the concept of stationarity in time series data, focusing on its defining traits such as constant mean and variance over time. Understand how to visually assess stationarity and apply the Augmented Dickey-Fuller test to statistically confirm whether your data meets this key assumption for forecasting models.
We'll cover the following...
Understanding stationarity
Stationarity is a property of time series data that indicates the absence of change in the mean and in the variance over time. A stationary time series will be horizontal, meaning it has no trend, and it will change always within a certain range.
Stationarity is an assumption used by some forecasting methods, so it's important to know how to identify it before applying those methods.
Identifying stationarity
Although there are statistical methods to identify stationarity, it is usually easy to visually inspect time series data to see if it's stationary or not.
Let's first try the visual method to have a better ...