(*) Doctorante à l'Université de Florence
Inverse projection, introduced by R.Lee in 1974, is a reconstruction technique of past populations. This technique uses an initial population age distribution and a given sequence of births and deaths to derive sequence of vital rates and age distributions. Lees Inverse Projection is a macrosimulation technique which is based on a deterministic methodology. The Stochastic Inverse Projection, elaborated by S. Bertino and E. Sonnino (Università " La Sapienza " of Roma), is a new procedure which reconstructs the past populations starting always from the series of births and deaths as the traditional Inverse Projection does through a microsimulation method based on the theory of the unhomogeneous Poisson process. Following the deterministic approach, there is just one reconstruction that is compatible with the input (data and assumptions). Instead, following the stochastic approach we obtain a set of different reconstructions for the same time t. The results of these simulations differ by chance only. These results are equally coherent because they are governed by the same laws and data given on input; so, the average of the data reconstructed offers a good idea of what might have happened in the past. In this talk, the methodology of the Stochastic Inverse Projection and his pecularities are illustrated. Moreover, a comparison between Inverse Projection and Stochastic Inverse Projection is shown by applying them to the Tuscan data.