Linear Exponential Smoothing Response Grid

Conceptual Overview: Explore how the two parameters of Holt's Linear Exponential Smoothing procedure influence the forecasts using a response grid.

Holt's Linear Exponential Smoothing is a generalization of exponential smoothing. It uses two smoothing constants: $\backslash$alpha for the level of the time series and $\backslash$beta for the slope of the time series. In this example of predicting bicycle sales, $\backslash$alpha $=0\mathrm{.}5$ and $\backslash$beta $=0\mathrm{.}5$ produce a mean squared error $($MSE$)$ of $7\mathrm{.}53.$ Drag the dot in the control grid below the graph to change both $\backslash$alpha and $\backslash$beta to search for combinations of those two parameters that would produce the lowest MSE.