Performing time series data decomposition
Decomposition is the process of splitting time series data into individual components to gain better insights into underlying patterns. In general, decomposition helps us to understand the underlying patterns in our time series better. The components are defined as follows:
- Trend: The long-term increase or decrease of the values in the time series
- Seasonality: The variations in the time series which are influenced by seasonal factors (e.g., quarter, month, week, or day)
- Residual: The patterns left after trend and seasonality have been accounted for. It is also considered noise (random variation in the time series)
As you may have noticed, cyclical variations covered in previous recipes do not appear as a component in decomposed time series. It is usually combined with the trend component and called trend.
When decomposing our time series, we can consider the time series as either an additive or multiplicative combination...
