Grant: Discovery Grants Program - Individual
University of Alberta
$19,000.00
The one-step approach to minimum distance estimation with applications
EF 2009-10
Résumé
The main objective of my proposed research program is to investigate statistical estimation with minimum distance techniques. Roughly speaking, minimum distance estimators are parameter estimates chosen to minimize a distance between the data and the fitted model. It has been observed that minimum distance estimators are automatically robust, which makes them appealing in practice. Several minimum distance estimators for various models have been considered in the literature, primarily based on distribution functions. In general, different distances produce estimators that are consistent and asymptotically normal. Computational feasibility is an important factor for any estimation procedure. In most cases, the ease of computing depends on the choice of the distance and whether the model is parametric or semi-parametric in nature.
Estimating parameters using the quadratic distance measure is the easiest, and most of the published results have focused on this problem. A more advanced distance measure that produces robust and efficient estimators is the density-based Hellinger distance measure. However, the usefulness of the preceding method is hampered by its intensive computation. This problem is more severe in semi-parametric settings than in parametric models due in large part to the fact that the minimum Hellinger distance estimation approach may require the minimization of a non-differentiable function. Clearly, it is of theoretical and practical interest to have a data-driven distance method that has computational expediency, robustness and high statistical efficiency. My research program will include investigation of the "one-step" approach to minimum distance estimation methods as a simple way out of avoiding the computational burden. For a wide range of models, I am planning to study one-step minimum distance estimators and their asymptotic properties.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Alberta, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
7987-2009
Name nom
Karunamuni, Rohana
Department département
Mathematical and Statistical Sciences
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2009
Competition year année de concours
2009
Award amount
19000
Program id
RGPIN
Program nane en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
14
Committee name en
Statistical Sciences
Committee name fr
Statistique
Area of application code
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
3001
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Statistical theory
Research subject fr
Théorie de statistique
Partie
1
Nb partie
5
Application title
The one-step approach to minimum distance estimation with applications
Keyword
NA
Application summary
The main objective of my proposed research program is to investigate statistical estimation with minimum distance techniques. Roughly speaking, minimum distance estimators are parameter estimates chosen to minimize a distance between the data and the fitted model. It has been observed that minimum distance estimators are automatically robust, which makes them appealing in practice. Several minimum distance estimators for various models have been considered in the literature, primarily based on distribution functions. In general, different distances produce estimators that are consistent and asymptotically normal. Computational feasibility is an important factor for any estimation procedure. In most cases, the ease of computing depends on the choice of the distance and whether the model is parametric or semi-parametric in nature.
Estimating parameters using the quadratic distance measure is the easiest, and most of the published results have focused on this problem. A more advanced distance measure that produces robust and efficient estimators is the density-based Hellinger distance measure. However, the usefulness of the preceding method is hampered by its intensive computation. This problem is more severe in semi-parametric settings than in parametric models due in large part to the fact that the minimum Hellinger distance estimation approach may require the minimization of a non-differentiable function. Clearly, it is of theoretical and practical interest to have a data-driven distance method that has computational expediency, robustness and high statistical efficiency. My research program will include investigation of the "one-step" approach to minimum distance estimation methods as a simple way out of avoiding the computational burden. For a wide range of models, I am planning to study one-step minimum distance estimators and their asymptotic properties.