Combination of parametric and nonparametric estimators for population abundance using line transect sampling

Abstract

This paper introduced a modification of the classical kernel estimator for population abundance using line transect sampling. The procedure allows the incorporation of a small set of parametric models with the usual kernel estimator. Akaike Information Criterion (AIC) is used to choose the most appropriate parametric detection function. The resultant estimators of f(0); the probability density function at perpendicular distance x = 0, are compared with the existing usual kernel estimator by using the simulation technique and a set of real data. Numerical results demonstrate the superiority of these estimators over the usual kernel estimator for almost all considered cases.