Bayesian inference for multinomial probabilities from a stopping time sample information
We undertake Bayesian inference of multinomial probabilities associated with a finite alphabet, under incomplete experimental information. Specifically, we observe the stopping time variable representing the number of letters needed to see a particular pattern for the first time. The procedure is extended to the case where for two fixed words, one appeared before the other. An application of the method to a reliability problem will be shown.
Keywords: Patterns stopping times incomplete experimental information exact Bayesian inference