I present an algorithm for inverting the luminosity function for white dwarfs to obtain a maximum likelihood estimate of the star formation rate of the host stellar population .
The algorithm is of the general class of Expectation Maximization , and involves iteratively improving an initial guess of the star formation rate .
Tests show that the inversion results are quite sensitive to the assumed metallicity and initial mass function , but relatively insensitive to the initial-final mass relation and ratio of H/He atmosphere white dwarfs .
Application to two independent determinations of the Solar neighbourhood white dwarf luminosity function gives similar results : the star formation rate is characterised by an early burst , and more recent peak at 2-3 Gyr in the past .