Title: Bayesian inference and prediction for high-frequency data using particle filtering
Authors: Jim Griffin - University of Kent (United Kingdom)
Abstract: Financial prices are usually modelled as continuous, often involving geometric Brownian motion with drift, leverage and possibly jump components. An alternative modelling approach allows financial observations to take discrete values when they are interpreted as integer multiples of a fixed quantity, the ticksize, the monetary value associated with a single change in the asset evolution. These samples are usually collected at very high frequency, exhibiting diverse trading operations in seconds. In this context, the observables are modelled via the Skellam process, defined as the difference between two independent Poisson processes. The intensities of the Poisson processes are modelled as functions of a stochastic volatility process, which is in turn described by a discretised Ornstein-Uhlenbeck AR(1) process. The choice of a discrete-time/discrete-space model is able to provide good fitting and out-of-sample prediction results, the latter being computed using Particle Filtering methods.