Analysis of the probability distribution method for spectrum decomposition：（频谱分解的概率分布的分析方法）.pdfVIP

Mikrochim.Acta 126,77-81 (1997) Mikrochimica Acta 9 Springer-Verlag1997 Printed in Austria Analysis of the Probability Distribution Method for Spectrum Decomposition Zsolt R6vay Department ofNuclear Physics,Institute of Isotopes,Hungarian Academyof Sciences,P.O. Box.77, H-1525Budapest,Hungary Abstract. The probability distribution method provides and extracts the most information. This is the a new way for spectrum deconvolution, that is probability distribution method (PDM), introduced especially applicable in the case of poor counting a few years ago [1]. P D M presumes a binomial statistics. Here it has been compared with the conven- distribution of counts in the spectrum channels. tional least-squares curve-fitting method. A large Methods based on the maximum likelihood principle number of spectrum regions of a few channels were may also apply the binomial distribution (instead of generated to simulate a spectrum containing a low the typically used Poisson distribution), but these intensity peak on low background, and were evaluated methods extract only a part of the information hidden by both methods, to establish which method provides in the spectrum. an unbiased estimate of the peak area in the cases The main difference of P D M from other methods is examined, and on what conditions.