Philipp di Dio
Universität Konstanz
Gaussian Mixtures from Derivatives of Moment Functionals to the Heat Equation
Gaussian Mixtures play an important role in algebraic statistics and data analysis. Despite their importance, theoretical knowledge and reconstruction results are limited. In the present talk we introduce the concept of derivatives of moments and moment functionals to reconstruct probability distributions. In a second step we extend this to the heat equation. The heat equation is the natural setting for Gaussians since the fundamental solution of is the Gaussian.
