Publisher Summary The maximum entropy method (MEM) appears to be a powerful technique to analyze relaxation data. The method is robust against experimental constraints such as the signal to noise ratio and signal truncation. This method needs no assumptions for initial values and number of components. The analysis of the relaxation curve as a continuous distribution with MEM gives more realistic physical results, which take into account the heterogeneity of most food samples. The analysis of multiexponential decay curves is the first step in the interpretation of the nuclear magnetic resonance (NMR) relaxation data. The second step is the interpretation of the relaxation parameters on the basis of the physical model, which should be proposed according to the physical or chemical properties of the sample. However, in the literature the physical model has often been adjusted to the result of data analysis which unfortunately depends on the numerical method used to fit the data. To avoid this problem, it is essential to test several numerical fitting methods before deciding the most realistic solution.
Mariette, F., Guillement, J. P., Tellier, C., & Marchal, P. (1996). Chapter 10 Continuous relaxation time distribution decomposition by mem. Data Handling in Science and Technology, 18(C), 218–234. https://doi.org/10.1016/S0922-3487(96)80047-5