A Parameter Estimation Method of Shock Model Constructed with Phase-Type Distribution on the Condition of Interval Data

Abstract EN

The phase-type distribution (also known as PH distribution) has mathematical properties of denseness and closure in calculation and is, therefore, widely used in shock model constructions describing occurrence time of a shock or its damage.

However, in the case of samples with only interval data, modeling with PH distribution will cause decoupling issues in parameter estimation.

Aiming at this problem, an approximate parameter estimation method based on building PH distribution with dynamic order is proposed.

Firstly, the shock model established by PH distribution and the likelihood function under samples with only interval data are briefly introduced.

Then, the principle and steps of the method are introduced in detail, and the derivation processes of some related formulas are also given.

Finally, the performance of the algorithm is illustrated by a case with three different types of distributions.