Pair-copula constructions for probabilistic networks

Abstract

Probabilistic Networks (PNs), also known as Bayesian networks, causal networks, belief networks or probabilistic dependence graphics, emerged in the 80's and have been applied in a wide variety of real-world activities. However, the most common available structure estimation algorithms underlying to probabilistic network classiers are often conned to discrete or Gaussian models. At the same time, copula constructions often lead to signicant improvement in density estimation in order to represent multivariate continuous distribution and Pair Copula Construction (PCC) presents a general construction method for multivariate copulas using only bivariate copulas, which allows an efective access to high-dimensional problems. In this project, we propose to investigate the conjunction of PCC to PN directed to machine learning approach in the task classification applied to real financial, industrial and medical data. (AU)