CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Adaptive estimation for mode and anti-mode preserving distribution on the circle Authors:  Takayuki Shiohama - Tokyo University of Science (Japan) [presenting]
Abstract: The efficient estimation for the parameters in mode and anti-mode preserving distribution on the circle is considered. For practical purpose, it is difficult to assume that the base density function in advance, which are transformed to be skewed and sharply peaked distributions by inverse functions. However, starting with an initial computationally convenient estimator, one can always gain efficiency by doing a one-step from an initial estimate using a likelihood or other efficient criterion function. The proposed methods are extended to the case with circular-circular regression models with mode and anti-mode preserving error distributions. Some Monte Carlo simulations are illustrated for the performance comparison of our proposed one-step estimators among other estimators.