A transformation for time-series data of the form
where xt is the transformed series, yt is the original series, and λ is a number between 0 and 1. It is often used to stabilize the variance of the time series. λ = 0 corresponds to the logarithmic transformation, and λ = 1 is a mean reduction. The optimal λ can be estimated from the data, simultaneously with other parameters in a completely specified time series model for yt.