astsa (version 1.4)

SVfilter: Switching Filter (for Stochastic Volatility Models)

Description

Performs a special case switching filter when the observational noise is a certain mixture of normals. Used to fit a stochastic volatility model.

Usage

SVfilter(num, y, phi0, phi1, sQ, alpha, sR0, mu1, sR1)

Arguments

num

number of observations

y

time series of returns

phi0

state constant

phi1

state transition parameter

sQ

state standard deviation

alpha

observation constant

sR0

observation error standard deviation for mixture component zero

mu1

observation error mean for mixture component one

sR1

observation error standard deviation for mixture component one

Value

xp

one-step-ahead prediction of the volatility

Pp

mean square prediction error of the volatility

like

the negative of the log likelihood at the given parameter values

References

http://www.stat.pitt.edu/stoffer/tsa4/