In this method, too, we start from an `easy' function , which `we hope' will approximate of interest, of which in fact we know only its un-normalized expression . However, there is no requirement about how approximates (but the goodness of approximation will have an impact on the efficacy of the method), apart from the condition that must be positive wherever is positive .

The function
can be used
in the calculation of
, if we
notice that
can be rewritten as follows:

where the the symbol is a reminder that the expectation is calculated over the distribution . Finally, the strategy can be implemented in the Monte Carlo using Eq. (111) for the two expectations:

(115) |

From the same sample it is also possible to evaluate the normalization constant, given by the denominator of Eq. (113), i.e.

(116) |

The computation of this quantity is particularly important when we are dealing with model comparison and has the meaning of `evidence' (Sect. 7).

It easily to see that the method works well if overlaps well with . Thus, a proper choice of can be made by studying where the probability mass of is concentrated (for example finding the mode of the distribution in a numerical way). Often a Gaussian function is used for , with parameters chosen to approximate in the proximity of the mode, as described in Sect. 5.10. In other cases, other functions can be used which have more pronounced tails, like -Student or Cauchy distributions. Special techniques, into which we cannot enter here, allow independent random numbers to be generated and, subsequently, by proper rotations, turned into other numbers which have a correlation matrix equal to that of the multi-dimensional Gaussian which approximates .

Note, finally, that, contrary to the rejection sampling, importance sampling is not suitable for generate samples of `unweighted events', such as those routinely used in the planning and the analysis of many kind experiments, especially particle physics experiments.