hestonMLE: Quasi MLE for Heston dynamics

In shill1729/findistr: Common distributions used in finance

Quasi MLE for Heston dynamics

Description

An ad-hoc step-wise QMLE routine for estimating parameters for the Heston dynamics given an observational time-series of log-price increments.

Usage

hestonMLE(y, iterations = 1, N = 100, nthresh = 100, h = 1/252)

Arguments

y	time-series of log-price increments
iterations	number of iterations to step through for filtering and
N	number of particles to use in the particle-filter
nthresh	the threshold of particles in the particle-filter
h	the time-step performing the MLE estimates

Details

At each step, hidden volatility is filtered into an estimate based on log-price observations, and conditional on these observations, the likelihood is maximized over the parameter space. Then on the next step, the volatility is re-filtered under the updated parameters until reaching the last observation. The filtered volatility is tracked step-wise, so that each iteration produces a new filtered-state appended to the previous filtered-states produced under the previous parameters.

Value

vector/numeric

shill1729/findistr documentation built on May 20, 2024, 9:43 a.m.