2.2.3.5 Decomposition

TSA-TD-Decomposition

Introduction

Decomposition is a technique used to separate a time series into several distinct components, making it easier to understand and model the underlying patterns. The main components typically include:

1. Trend Component: The long-term progression of the series, showing the overall direction over a longer period (e.g., upward or downward movement).
2. Seasonal Component: The repeating short-term cycle in the series, reflecting regular periodic fluctuations (e.g., monthly or quarterly patterns).
3. Residual Component: The random noise or irregular variations that are not explained by the trend or seasonal components.

Decomposition helps in understanding the underlying structure of the time series data by isolating these components, which can then be analyzed separately. This is particularly useful for forecasting and identifying the different influences on the data. There are two main types of decomposition:

1. Multiplicative: Used when the components multiply together to form the time series.
2. Additive: Used when the components add together to form the time series.

More information:

• Tutorial: Decomposition
• Algorithm: Decomposition