Title :
On the midrange estimator
Author :
Arce, Gonzalo R. ; Fontana, Sharon A.
Author_Institution :
Dept. of Electr. Eng., Delaware Univ., Newark, DE, USA
fDate :
6/1/1988 12:00:00 AM
Abstract :
An error criterion for the design of an estimator based on extreme-order statistics, namely, the midrange, is defined. This approach is taken to overcome the fact that the variance of the midrange estimate is strongly dependent on the input distribution. Moreover, under this performance criterion, nonparametric characteristics for the design of these estimators are found
Keywords :
estimation theory; nonparametric statistics; error criterion; extreme-order statistics; input distribution; midrange estimator; nonparametric characteristics; variance; Acoustics; Additive white noise; Density functional theory; Filters; Gaussian distribution; Limiting; Power distribution; Probability distribution; Signal processing; Speech;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
