SMS scnews item created by John Robinson at Fri 10 Oct 2008 1557
Type: Seminar
Distribution: World
Expiry: 17 Oct 2008
Calendar1: 17 Oct 2008 1400-1500
CalLoc1: Carslaw 173
Auth: johnr(.ststaff;3005.3001)@p8224.pc.maths.usyd.edu.au

Statistics Seminar: Peiris -- Estimation of Duration Models in Finance

Estimation of Duration Models in Finance: Semiparametric and QMLE Methods Shelton Peiris
School of Mathematics and Statistics The University of Sydney, NSW 2006 

Abstract: 

We consider a class of time series models to analyze the time gaps between transactions
(durations) called Autoregressive Conditional Duration (ACD) models.  Since the MLE
procedure is difficult to implement, we suggest a semiparametric estimation method based
on estimating functions.We provide an example based on a real data set to illustrate
this new approach.  We also discuss the class of log ACD models and use QMLE methods for
parameter estimation following Allen et al (2008), Journal of Econometrics.  Examples
from real data sets are provided to illustrate this QMLE estimation if time permits.  

Key words: Autoregressive, Conditional expectation, Intensity, Hazard function,
Stochastic process, Prediction, Estimation, Irregular data, Transaction data, Finance,
Durations, Autocorrelations, Volatility.