The Smearing retransformation is used in regression analysis, after estimating the logarithm of a variable. Estimating the logarithm of a variable instead of the variable itself is a common technique to more closely approximate normality. In order to retransform the variable back to level from log, the Smearing retransformation is used.
If the log-transformed variable y is normally distributed with mean
f(X)
\sigma2
then, the expected value of y is given by:
y=\exp(f(X))\exp(
1 | |
2 |
\sigma2).