Skip to main content
European Commission logo print header
Zawartość zarchiwizowana w dniu 2022-12-23

Estimation of parameters and construction of goodness-of-fit tests in some non- and semi-parametric models

Cel

There has been a dramatic change in the topics of theoretical research within mathematical statistics. Classical statistics considers finite dimensional parametric models. These are the widely used standard models for many fields of application like econometrics, psychometrics, chemometrics. Nowadays much research is going on for far more realistic models with large parameter spaces, and this has led to new theories
for inference in so-called semi- and non-parametric models. Within this branch of semi- and non-parametric statistics, the proposed research will focus on estimation for the important and broad class of mixture models and on estimation and goodness-of-fit testing for models with incomplete data. Mixture models as well as incomplete data models are applied very frequently in practice.

The expert knowledge available in the teams from Tbilisi, Diepenbeek and Amsterdam is well-suited for combined efforts in research in this field. Moreover, the proposed mix of old and new links between the researchers involved will undoubtedly contribute to the success of the project. The research of these three teams in this field will be facilitated by intensive e-mail contacts and be stimulated enormously by mutual work-visits. In this way, the proposed research will lead to quite some publications and progress in this field.

The proposal categorizes the research to be done into five subtasks or objectives as follows (see 3.1.4.1 ):

1. To construct estimators of the mixing distribution in certain mixture models and to obtain optimal convergence rates for estimation in those models.
2. To investigate the problem of estimating mixing distributions when observations are censored.
3. To derive estimators in semi-parametric mixture models and to investigate the asymptotic properties of the proposed estimators.
4. To investigate the asymptotic properties (consistency, rates of convergence, efficiency) of estimators in semi-parametric change-set problems.
5. To construct non-parametric estimators and goodness-of-fit tests in models with incomplete data (censored, truncated).

Zaproszenie do składania wniosków

Data not available

System finansowania

Data not available

Koordynator

University of Amsterdam
Wkład UE
Brak danych
Adres
Plantage Muidergracht 24
1018 TV Amsterdam
Niderlandy

Zobacz na mapie

Koszt całkowity
Brak danych

Uczestnicy (2)