ALGORITHMS OF NUCLIDES ACTIVITIES ESTIMATION BY THE 
MAXIMUM LIKELIHOOD METHOD
V. A. Muravsky, S. A. Tolstov
International Sakharov Institute of Radioecology
23, Dolgobrodskaya Str., 220009 Minsk, Belarus
     Among the methods used for estimating of the activity of nuclides
on  spectra  of  ionizing  radiation,  the  maximum  likelihood method
(ML-method) has the best accuracy. However, they often prefer to apply
more  simple but less precise methods,  such as the least squares.  At
the same time,  the ML-method allows to increase activities  estimates
accuracy  essentially  (up to two - ten times for low activities) [1].
Therefore, the problem of development on the basis of the ML-method of
the algorithms of activities estimation is actual.
     One should  solve  a  number  of problems,  when one develops the
algorithms on the basis of the ML-method:
     - how to take into account errors in calibration data;
     - how to take into account the instability of the  spectrometer's
energy scale;
     - how to take into account density of the sample;
     - how to test the adequacy of the spectrum theoretical model.
     A way of solution of these problems is considered in the proposed
work  on  an example of the based on the ML-method nuclides activities
estimation algorithms developed in International Sakharov Institute of
Radioecology  and  realized  in  a  number  of  computer  programs for
scintillation gamma- and beta-spectrometers.
     References
     1. V.A.  Muravsky,  S.A.  Tolstov,  Sravneniye  metodov   otsenki
component   v   spectrakh  nizkoi  intensivnosti.  (Zurnal  Prikladnoi
Spectrometrii) /  Comparison  of  methods  for  low-intensity  spectra
components estimation (Journal of Applied Spectrometry),  v. 65, N. 3,
1998, 346-349, in Russian.

Back

1