A general method is proposed to construct fuzzy tests for testing statistical hypotheses about an unknown fuzzy parameter, when the data available are observations of a fuzzy random variable. The proposed method to construct such tests is essentially based on the one-to-one correspondence between the acceptance region of any level alpha test and the 1-alpha level confidence interval for the same parameter. By using this equivalency in the case of fuzzy environment, we construct the fuzzy test showing the degree of acceptability of the null and alternative hypotheses of interest. A numerical example in the field of lifetime study is given to clarify the theoretical results.

Keywords