Please use this identifier to cite or link to this item:
DC FieldValue
dc.titleScaling, clustering and dynamics of volatility in financial time series
dc.contributor.authorYUAN BAOSHENG
dc.identifier.citationYUAN BAOSHENG (2006-05-24). Scaling, clustering and dynamics of volatility in financial time series. ScholarBank@NUS Repository.
dc.description.abstractThis thesis investigates scaling, clustering and dynamics of volatility in financial time series (FTS) and studies the underlying mechanism. We propose a direct measure of volatility clustering (VC) based on the conditional probability distribution (CPD) of returns of FTS; We demonstrate that the CPDs of real FTS exhibits universal scaling, characterized by a collapse of the CPDs into to a universal curve with a power-law tail. We construct a simple phenomenological model to explain the emergence of VC and the associated volatility scaling. We introduce dynamical risk aversion (DRA) to study the impact of heterogeneous agents on the price dynamics; and we show that the DRA is the primary driving force responsible for excess price fluctuations and the associated VC. We lastly investigate herding behavior in financial markets in the context of an evolutionary Minority Game; we discover a general mechanism for the transition of agents from segregation into opposing groups to clustering towards cautious behavior. All these results may shed some light in understanding the dynamics of FTS and their underlying mechanism.
dc.subjectScaling, volatility clusteing, dynamics, Financial Time Seires, Conditional probability measure
dc.contributor.supervisorCHEN KAN
dc.description.degreeconferredDOCTOR OF PHILOSOPHY
Appears in Collections:Ph.D Theses (Open)

Show simple item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Yuan Baosheng PhD Thesis Final.pdf10.16 MBAdobe PDF



Page view(s)

checked on Jun 14, 2019


checked on Jun 14, 2019

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.