Artificial neural network as an effective tool to calculate parameters of positron annihilation lifetime spectra

The paper presents the application of the multi-layer perceptron regressor model for predicting the parameters of positron annihilation lifetime spectra using the example of alkanes in the solid phase. Good agreement of calculation results was found when the approach is compared with the commonly used methods, e.g., LT. The presented method can be used as an alternative quick and accurate tool for the decomposition of positron annihilation lifetime spectroscopy (PALS) spectra in general. The advantages and disadvantages of this new method are discussed. We show the preliminary results where the trained network can give better outcomes than the results yielded by programs based on an analysis of a single PALS spectrum.