Seminar in applied mathematics and statistics

SPEAKER: Benjamin Christoffersen (Copenhagen Business School).

TITLE: Particle smoothers in hidden Markov models for modeling financial distress.

ABSTRACT:  Discrete hazard models have become standard models in probability distress models for corporations. A major drawback of these models is that they often do not capture the clustering of distresses in certain periods which is the periods where lenders have the biggest losses. Hence, the models are extended to hidden Markov models to account for the excess clustering. However, this extension complicates the estimation as it requires integration over a high dimensional space. I show two recent particle smoothers which can be used in the E-step of an EM algorithm to compute the expected log-likelihood. I cover the pros and cons of both methods and the implications of various choices when implementing the particle filters and smoothers.

Tea and chocolate will be served in room 04.4.19 after the seminar.

----------

Upcoming events:

Wednesday, November 28 at 15.15: Shota Katayama

Wednesday, December 5 at 14.15: Rafael Serrano

Friday, December 14 at 14.15: Moritz M. Schauer

Friday, February 8, 2019, at 14.15: Massimiliano Tamborrino

Friday, February 15, 2019 at 14.15: Irene Tubikanec