The term "smoothing" is often used to refer to techniques that can be applied to time series data in order to produce smoothed (less noisy or slower moving) data for presentation, or to make out-of-sample forecasts.
The techniques supported here range from a simple weighted-moving average (WMAi) to exponential smoothing algorithms. The WMA averages the fixed number of past observations with fixed weights, while the exponential smoothing assigns exponentially decreasing weights over time, including all past observations.
The smoothing techniques vary in their complexity, based on how they handle trend and seasonality in the time series:
- Weighted-moving average (WMA) and Holt's simple exponential smoothing assume a stationary time series
- Holt-Winters double exponential smoothing and Holt's linear exponential smoothing are ideal for time series that possess a trend
- Winters's triple exponential smoothing takes into account seasonal changes as well as trends.
- Exponential smoothing is commonly applied to financial market and economic data, but it can be used with any discrete set of repeated measurements.