Univariate parametric parzen approximation using an efficient frequency domain approach
This paper introduces a frequency domain approximation to the univariate Parzen probability density function estimator. It is applied to construct an efficient semi-parametric density function estimator and density function distance measure that operates directly from sample data. The complexity of the techniques are constant with respect to the number of data samples and only depends on a scalar factor controlling the accuracy of the estimate. Copyright © 2002 IEEE.