An algorithm for constructing high dimensional distributions from distributions of lower dimension

Citation:

Anatolyev, Stanislav, Renat Khabibullin and Artem Prokhorov (2014) "An algorithm for constructing high dimensional distributions from distributions of lower dimension", Economics Letters, vol. 123, no. 3, pp. 257-261

Abstract:

We propose a new sequential procedure for estimating multivariate distributions in cases when conventional maximum likelihood has too many parameters and is therefore inaccurate or non-operational. The procedure constructs a multivariate distribution and its pseudo-likelihood sequentially, in each step using lower-dimensional distributions with a small number of parameters. In an application, the procedure provides excellent fit when the dimension is moderate, and remains operational when the conventional method fails.

Paper in RePEc:

Economics Letters, Vol. 123, pp. 257-261

Paper in accepted version:

RecDD.pdf

Supplementary material:

Extended paper with MATLAB codes