Critical evaluation of goodness of fit of computerised glow curve deconvolution

Abstract

Computerised glow curve deconvolution (CGCD) technique is a powerful tool for evaluation of trapping parameters of the thermoluminescence glow curve. A difficulty in the outcome is the level of confidence in accepting the values of the parameters that the technique yields. At this juncture the importance of goodness of fit test arises to compare the deviations of experimental from the expected values obtained from the model. As far as dosimetric application is concerned the leaders of the field have propagated the Figure of Merit (FOM) as criteria to test the goodness of fit. In this paper, three well accepted conventional methods of goodness of fit statistics namely Discrepancies (DIS), Reduced Chi-square and Chi-square (<img src='/image/spc_char/lembda.gif' border=0>²) test are used with FOM to critically evaluate the potentialities of the statistical tests. Glow curves of iodised salt are taken as reference glow curves.