Abstract
A class of nonparametric mode estimators is proposed. While the widely applied half sample mode estimators use the diameter of a set as the “measure of concentration”, the proposed estimators use for it some types of “variance measures”. In some cases, the new estimators perform better than half sample mode and half range mode estimators. Strong consistency is proved for an estimator from the class.
Original language | English |
---|---|
Number of pages | 14 |
Journal | Communications in Statistics: Simulation and Computation |
DOIs | |
Publication status | Published - 22 Jan 2020 |
Keywords
- Fraction-of-sample mode
- Half sample mode
- Nonparametric mode estimator
- ROBUST ESTIMATORS
- DENSITY-FUNCTION
- MULTIVARIATE
- CONVERGENCE
- EFFICIENT ESTIMATION