Seminario Carlos N. Bouza - variazione

[Tommaso Proietti <proietti@dss.uniud.it>]

=================================================================

       Forum per la Societa' Italiana di Statistica

                       AVVERTENZA

     il nuovo servizio viene offerto in via sperimentale
     dal Dipartimento di Scienze Statistiche di Perugia
           con il software Majordomo 1.94.5
       istallato su una Workstation UNIX (AIX 4.3.3)

           Ci scusiamo per eventuali disguidi !

=================================================================


A correzione del precedente messaggio si avvisa che il seminario:

"Ranked Survey Sampling Modeling"

del Prof. Carlos N. Bouza

Departamento de Matemātica Aplicada
Facultad de Matemātica y Computaciōn
Universidad de La Habana, Cuba

avrā luogo presso il DIP. SCIENZE STATISTICHE, Saletta seminari
Via Treppo 18, UDINE

il giorno martedė 17 ottobre, alle ore 15.30 (anziché il giorno 10)  

Abstract
 
In different applications the statistician deals with the need  of
combining  control  and flexibility when surveying a population.  This is a
common problem in different studies.  For example if a  waste site is
studied the expert or decision maker [DM] is able to evaluate the randomly
selected different sampling sites and to establish a certain order by
evaluating subjectively the importance of each of them, by predicting the
value of the  variable of interest Y or by measuring a "cheap" one X.
Similar situation arises when a physician evaluates radiography results
device and ranks the  observed patterns (X) before deciding which patients
will constitute the sample to be evaluated by Axial Thomography.  To each
sampling unit a rank is assigned.  If m independent samples are selected,
by using Simple Random Sampling with replacement [SRSWR], and ranked using
X we can select the units taking into account the information given by the
ranks.  McIntyre [1952] proposed to use this information in the inference.
The method was named Ranked Set Sampling [RSS]. Recently this design has
been rediscovered and is generating a different alternative models. For
example See Patil et. al. [1994], Handbook of Statistics Vol. 12,  Stokes
[1995], Ann. Inst. Stat. Math., Kaur et. al. [1997], Biometrics, Samawi
[1996],  Muttlak [1998],  Bouza [submitted to] Biometrical Journal.
We  discuss the behavior of design based estimators of the population mean
m including the sample mean. Well known estimators are characterized:
ratio-type, difference , regression estimators.  They are compared in terms
of their accuracy.  The superpopulation approach is used as an alternative
and the solution  when non-responses are present is also studied.
Monte Carlo experiments permit to establish certain recommendations for
selecting the estimator to be used.
---------------------------------------------------------------------------

La S.V. č invitata a partecipare.
-------------------------------------------------
Tommaso Proietti                   
Dipartimento di Scienze Statistiche
Universitā di Udine
Via Treppo 18, 33100 Udine, Italy
Tel. +39 0432 24.9581
Fax  +39 0432 24.9595
------------------------------------------------