THE ROYAL STATISTICAL SOCIETY
=================================
TIMSAC (Time Series and Applications) Study Group Meeting
---------------------------------------------------------------
Thursday 17 June 1999, 4:00 pm
==========================================
The Royal Statistical Society, 12 Errol Street, London EC1
(Nearest Underground stations: Barbican, Moorgate)
CHRIS CHATFIELD (University of Bath)
"A new look at exponential smoothing "
JEREMY PENZER (London School of Economics)
" State space models for messy time series"
A L L A R E W E L C O M E !
SUMMARIES
CHRIS CHATFIELD
Many models, for which different forms of ES are optimal, are reviewed,
including ARIMA, state-space and non-linear models.
In particular, models for multiplicative Holt-Winters are given
which allow the variance to depend on the level and seasonality.
Some simple identification tools are given.
Results for simulated and real data are reported.
Tea break at 4:45 p.m.
JEREMY PENZER
Models for aberrant behaviour in times series are proposed. Our
method allows test statistics for many interventions to be computed from a
single run of the Kalman filter smoother. Departures, such as outliers and
level shifts, are readily dealt with. Extensions to seasonal and stochastic
volatility models are discussed.
Information on TIMSAC can be found on:
http://www.zoo.cam.ac.uk/timsac/timsac.htm
or contact:
[log in to unmask],
(Barbel Finkenstadt, University of Cambridge, Dept Zoology,
Downing Street, Cambridge CB2 3EJ, tel/fax ++44 (0)1223 336644)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|