Testing of robustness and efficiency of Rényi divergence estimators of probability densities

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21340%2F12%3A00199585" target="_blank" >RIV/68407700:21340/12:00199585 - isvavai.cz</a>
Result on the web
—
DOI - Digital Object Identifier
—

Result language
angličtina
Original language name
Testing of robustness and efficiency of Rényi divergence estimators of probability densities
Original language description
In this contribution we study R´enyi pseudo-distance estimators which are based on minimization of information-theoretic divergences between empirical and hypothetical probability distribution. These distances are more robust (than e.g. MLE estimators) against outliers and other measurement errors potentially present in the data sets. Robustness of these estimators is described by influence function. In [1] and [4] authors found explicit formulas for enumeration of Rényi pseudodistances in normal families and for their influence functions. We focus on finding explicit formulas for other families (Weibull, Cauchy, Exponential) and finding influence functions for these stimators. We perform computer simulations for pseudorandom contaminated and uncontaminated data sets, different sample sizes and different R´enyi pseudodistance parameters.
Czech name
—
Czech description
—

Project
<a href="/en/project/LG12020" target="_blank" >LG12020: Advanced statistical analysis and non-statistical separation techniques for physical processing detection in data sets sampled by means of elementary particle accelerators.</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach

Publication year
2012
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Similar results(10)