Show TOC

Function documentationTrend Dampening

 

For forecast models with a trend component (linear regression, the trend models for exponential smoothing with or without seasonality, and, in some cases, automatic model selection), you can dampen the trend for future forecast values by specifying a trend damping factor.

Use the trend dampening factor if you expect the past growth rate to slow down or intensify in future.

Features

The trend dampening factor is a number that is multiplied by the trend value (growth rate) when the forecast value is calculated. This makes it possible to slow down or intensify growth in a long-term trend.

  • With a trend dampening factor of less than 1, a type of saturation effect is produced. The trend value will exponentially converge to 0.

    Example Example

    A trend dampening factor of 0.9 decreases the growth rate for each period by 10% recursively.

    If the number of cars sold is currently growing by 1000 per period for example, the growth in the number of cars sold in the next period would be 0.9 * 1000 = 900, and would be 0.9 * 900 = 810 in the period after that.

    End of the example.

  • To achieve the reverse effect, you can enter a trend dampening factor greater than one.

Activities

To perform trend dampening, choose this option and specify a trend dampening factor in accordance with your expectations.