作者: Stephen W. Provencher
DOI: 10.1002/MACP.1979.021800119
关键词: Stability (probability) 、 Inverse distribution 、 Distribution (mathematics) 、 A priori and a posteriori 、 Polymer chemistry 、 Dispersity 、 Integral equation 、 Mathematics 、 Statistical physics 、 Inverse problem 、 Instability
摘要: A method is presented for inverting the Fredholm integral equations of first kind that arise in many experimental determinations molecular weight distributions high polymers. The well known problems instability and nonuniqueness are minimized by constraining distribution to be smoothest nonnegative one consistent with data. Analyses simulated photon correlation spectroscopy data show good accuracy can obtained unimodal fair certain bimodal ones, provided systematic errors kept below random noise level. completely automatic no initial estimate or prior information about needed. However, if there happens a priori knowledge it bimodal, accordingly further improve stability accuracy.