Comparison of the least squares and the maximum likelihood estimators for gamma-spectrometry

A comparison of the characteristics of the maximum likelihood (ML) and the least squares (LS) estimators of nuclides activities for low-intensity scintillation gamma-spectra has been carried out by the computer simulation. It has been found that the part of the LS estimators gives biased activity estimates and the bias grows with increase of the multichannel analyzer resolution (the number of the spectrum channels). Such bias in estimates leads to the significant deterioration of the estimation accuracy for low-intensity spectra. Consequently, the threshold of nuclides detection rises up to 2-10 times in comparison with the ML estimator. It has been shown that the ML estimator and the special LS estimator provide non biased estimates of nuclides activities. Thus, these estimators are optimal for practical application to low-intensity spectrometry. (C) 1998 Elsevier Science B.V. All rights reserved.