Particle MCMC for a time changed Ĺevy process

Yüksel, Ayhan
For almost any type of financial modelling exercise, the most fundamental problem is findingsuitablestochasticprocessesthatcapturetheobservedbehaviourofassetprices well. Stochastic volatility models, and their extensions with jumps, are class of flexible models that can capture such empirical dynamics quite well. However this richer modelling environment comes at the expense of estimation challenges. Estimation of these flexible models involves some additional challenges that do not exist for simpler ones. In this thesis, motivated by models with stochastic volatility and jumps, simulation based Bayesian inference methods are analyzed. First we discuss different Markov Chain Monte Carlo (MCMC) approaches in detail, develop various algorithms and implement them for a basic stochastic volatility model. Next we turn our attention to on-line inference and analyze particle filtering methods. We begin with a simple particle filter and then discuss methods to improve the basic filter. We also develop MonteCarloalgorithmstoimplementparticlefiltersforourstochasticvolatilitymodel. More advanced financial models typically include many latent random variables and complicated likelihood functions where standard MCMC methods may fail to efficientlyestimatethem. Asamoreeffectivealternative,wediscussedtheParticleMCMC methods recently proposed by C. Andrieu, A. Doucet, and R. Holenstein (Particle Markov Chain Monte Carlo. Journal of the Royal Statistical Society: Series B 72 (3), 2010, pp 269-342). Particle MCMC methods combine two strands of simulation based Bayesian inference, namely, particle filtering and MCMC, and offer a powerful tool for estimating complex financial models. The theoretical foundations for particle MCMC as well as various samplers proposed in the literature are analyzed in the thesis. In the final part of the thesis, we develop MCMC and particle MCMC methods for a stock price model with a time changed Ĺevy process. We assume that the stock pricefollowsaHeston-typestochasticvolatilityplusvariance-gammajumpsinreturns. Variance-Gamma process is an infinite activity finite variation Ĺevy process obtained bysubordinatinganarithmeticBrownianmotionwithaGammaprocess. Themodelis quiteflexibleinitsnatureandcancapturemostoftheobservedcharacteristicsofstock prices. Our main contribution to existing academic literature is the efficient particle MCMC algorithms that are developed for the Ĺevy based model. We compare MCMC and particle MCMC algorithms in an empirical implementation using S&P500 Index with 15 years of data. The results indicate that the particle MCMC algorithm is a more efficient alternative to standardMCMCandtypicallygivessmallerstandarderrorsand lower autocorrelations.


Comparison of Dictionary-Based Image Reconstruction Algorithms for Inverse Problems
Dogan, Didem; Öktem, Sevinç Figen (2020-10-07)
Many inverse problems in imaging involve measurements that are in the form of convolutions. Sparsity priors are widely exploited in their solutions for regularization as these problems are generally ill-posed. In this work, we develop image reconstruction methods for these inverse problems using patchbased and convolutional sparse models. The resulting regularized inverse problems are solved via the alternating direction method of multipliers (ADMM). The performance of the developed algorithms is investigat...
Shortcuts to spherically symmetric solutions: a cautionary note
Deser, S; Franklin, J; Tekin, Bayram (IOP Publishing, 2004-11-21)
Spherically symmetric solutions of generic gravitational models are optimally, and legitimately, obtained by expressing the action in terms of the surviving metric components. This shortcut is not to be overdone; however, a one-function ansatz invalidates it, as illustrated by the incorrect solutions of Wohlfarth (2004 Class. Quantum Grav. 21 1927).
Numerical simulation of rainfall-induced landslides considering hysteresis and scanning soil-water characteristic curves
Kenanoğlu, Melih Birhan; Toker, Nabi Kartal (2017-05-11)
Infiltration through unsaturated soil is essential issue in the modelling of rainfall-induced landslides, Hysteresis in soil-water characteristic curves complicates the modelling of infiltration. This study aims to establish a numerical model for rainfall infiltration and ultimate limit state stability of an unsaturated infinite slope, while considering the effect of wetting-drying hysteresis.
Nonlinear modelling and control of the flow over aerofoils using CFD simulations
Karaca, H. D.; Alıcı, Gökçen Deniz; KASNAKOĞLU, Coşku (2016-09-01)
A simulation based approach for nonlinear dynamical modelling and feedback control of the drag to lift ratio for aerofoils is investigated through case studies involving NACA 23012, ag13 and b737a aerofoils. The flow around the aerofoils is studied via numerical solutions of the 2D Navier-Stokes (NS) equations. A standard computational fluid dynamics (CFD) solver is extended to be able to measure desired feedback values and to apply a control input to the flow field. The proposed modelling and control appro...
Cartesian Grid Mapper: Transformation Media for Modeling Arbitrary Curved Boundaries With Cartesian Grids
ÖZGÜN, ÖZLEM; Kuzuoğlu, Mustafa (2014-01-01)
A transformation electromagnetics-based approach is presented to facilitate the use of fixed Cartesian grids for modeling arbitrarily shaped (convex or nonconvex) curved boundaries in finite methods. The basic idea is to design transformation media adapted to the Cartesian grid by using the duality between the material parameters of the media and the coordinate transformation that maps the points of the Cartesian grid to the points on the original curved boundary. Such a transformation medium is named Carte...
Citation Formats
A. Yüksel, “Particle MCMC for a time changed Ĺevy process,” Ph.D. - Doctoral Program, Middle East Technical University, 2015.