1. Advanced fitting algorithms for analysing positron annihilation lifetime spectra
- Author
-
Aurelia W. Dong, Ben J. Boyd, Calum J. Drummond, Anita J. Hill, Steven J. Pas, and C. Pascual-Izarra
- Subjects
Physics ,Complex data type ,Nuclear and High Energy Physics ,Markov chain ,Computer program ,Reliability (computer networking) ,Simulated annealing ,Experimental data ,Bayesian inference ,Focus (optics) ,Instrumentation ,Algorithm - Abstract
The most common way to analyse PALS spectra involves fitting a parameter-dependent model to the experimental data. Traditionally, this fit involves local non-linear optimisation routines that depend on a reasonable initial guess for the searched parameters. This, together with the fact that very different sets of parameters may yield indistinguishably good fits for a given experimental spectrum, gives rise to ambiguities in the data analysis in most but the simplest cases. In order to alleviate these difficulties, a computer program named PAScual was developed that incorporates 2 advanced algorithms to provide a robust fitting tool: on the one hand, it incorporates a global non-linear optimisation routine based on the Simulated Annealing algorithm and, on the other hand, it yields information on the reliability of the results by means of a Markov Chain Monte-Carlo Bayesian Inference method. In this work the methods used in PAScual are described and tested against both simulated and experimental spectra, comparing the results with those from the well-established program LTv9. The examples focus on the type of complex data that results from the study of self-assembled amphiphile materials containing co-existing aqueous and hydrocarbon regions.
- Published
- 2009
- Full Text
- View/download PDF