Neural Network Based Algorithm Of Preliminary Data Analysis: Application To Fluorescence And Esr Spectroscopy

Abstract

The fluorescence and electron spin resonance (ESR) spectroscopy are very important experimental tools for studying complex biomolecular objects and systems. The analysis of spectroscopic experimental data is often conducted by means of fitting using an analytical or a simulation models. For the successful performance of the fitting operation, an adequate model should be selected and good initial estimations of its parameters should be made. We propose to use artificial neural networks (ANNs) to recognise the appropriate model and to produce initial estimations of model's parameters before fitting.