In this paper we show that the standard definition of likelihoodfunction used in Bayesian inference simply and correctly handlesquantized measurements and imprecise likelihood functions. Somerecent papers have stated or implied that methods involving randomsets, fuzzy membership functions, generalized likelihood functions,or Dempster-Shafer concepts are required to handle impreciselikelihood functions and quantized measurements. While it is truethat one can use these methods, in the spirit of Occam’s razor, wefeel the simplest correct solution is the best.
展开▼