BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20181009T130925Z
DTSTART:20180529T140000Z
DTEND:20180529T153000Z
SUMMARY:CMI Afternoon Seminar: Grouped functional time series forecasting
  method for multiple sub-populations
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}u2n-jgxnnno
 m-qdjdgq
DESCRIPTION:Associate Professor Han Lin Shang\, Associate Professor of St
 atistics\, Australian National University\n\nAbstract: Age-specific mort
 ality rates are often disaggregated by different attributes\, such as se
 x\, state\,\nethnic group\, education and socioeconomic status. Forecast
 ing age-specific mortality rates at the\nnational and sub-national level
 s play a vital role in developing social policy and pricing annuity.\nHo
 wever\, the independent mortality forecasts at the sub-national levels m
 ay not add up to the\nforecasts at the national level. Further\, the ind
 ependent forecasts may not utilize correlation\namong sub-populations to
  improve forecast accuracy. To address these two issues\, we modify\nand
  extend the grouped univariate functional time series to grouped multiva
 riate functional time\nseries forecasting. For quantifying forecast unce
 rtainty\, we utilize a nonparametric bootstrap\nmethod to reconcile inte
 rval forecasts. Using the regional age-specific mortality rates in Japan
 \nobtained from the Japanese Mortality Database (2018)\, we investigate 
 the one-step-ahead to 15-\nstep-ahead forecast accuracy among the indepe
 ndent and grouped univariate and multivariate\nfunctional time series fo
 recasting methods. The grouped multivariate functional time series\nfore
 casting methods are not only shown to be useful for reconciling forecast
 s of age-specific\nmortality rates at national and sub-national levels\,
  but they also use multivariate functional\nprincipal component regressi
 on to jointly model sub-populations and enjoy potentially improved\nfore
 cast accuracy averaged over different disaggregation factors. The improv
 ed forecast accuracy\nof mortality rates is of great interest to the ins
 urance and pension industries for estimating\nannuity prices\, in partic
 ular at the level of population sub-groups\, defined by critical factors
 \nsuch as sex\, region\, and socioeconomic grouping.\n\nTea/coffee and c
 akes from 2.45.\n\nJoin us for this event\, which is part of the CMI Aft
 ernoon Seminar Series.  All welcome.  No registration necessary.\n\nThe 
 Cathie Marsh Institute for Social Research (CMI) provides a focal point 
 at The University of Manchester for the application of quantitative meth
 ods in interdisciplinary social science research in order to generate a 
 world class research environment.\n
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:CMI Seminar Room\, 2.07\, Humanities Bridgeford Street\, Manches
 ter
END:VEVENT
END:VCALENDAR