Abstract
Heavy-tailed distributions, prevalent in a lot of real-world applicationssuch as finance, telecommunications, queuing theory, and natural languageprocessing, are challenging to model accurately owing to their slow tail decay.Bernstein phase-type (BPH) distributions, through their analytical tractabilityand good approximations in the non-tail region, can present a good solution,but they suffer from an inability to reproduce these heavy-tailed behaviorsexactly, thus leading to inadequate performance in important tail areas. On thecontrary, while highly adaptable to heavy-tailed distributions,hyperexponential (HE) models struggle in the body part of the distribution.Additionally, they are highly sensitive to initial parameter selection,significantly affecting their precision. To solve these issues, we propose a novel hybrid model of BPH and HEdistributions, borrowing the most desirable features from each for enhancedapproximation quality. Specifically, we leverage an optimization to set initialparameters for the HE component, significantly enhancing its robustness andreducing the possibility that the associated procedure results in an invalid HEmodel. Experimental validation demonstrates that the novel hybrid approach ismore performant than individual application of BPH or HE models. Moreprecisely, it can capture both the body and the tail of heavy-taileddistributions, with a considerable enhancement in matching parameters such asmean and coefficient of variation. Additional validation through experimentsutilizing queuing theory proves the practical usefulness, accuracy, andprecision of our hybrid approach.