Dynamic Adaptive Mixture Models

TL;DR

This paper introduces Dynamic Adaptive Mixture Models (DAMMs) that adapt mixture components and weights in real-time, enabling exact likelihood computation and efficient approximation of complex models, with applications demonstrated in financial econometrics.

Contribution

The paper presents a novel class of models that adapt mixture components and weights sequentially, avoiding intensive simulations and enabling exact likelihood calculation.

Findings

DAMMs accurately approximate complex stochastic mixture models

The models can adapt to data in real-time

Applications show effectiveness in financial econometrics

Abstract

In this paper we propose a new class of Dynamic Mixture Models (DAMMs) being able to sequentially adapt the mixture components as well as the mixture composition using information coming from the data. The information driven nature of the proposed class of models allows to exactly compute the full likelihood and to avoid computer intensive simulation schemes. An extensive Monte Carlo experiment reveals that the new proposed model can accurately approximate the more complicated Stochastic Dynamic Mixture Model previously introduced in the literature as well as other kind of models. The properties of the new proposed class of models are discussed through the paper and an application in financial econometrics is reported.