The Study On Bootstrap Methods For Fuzzy Hypotheses Testing

The Fuzzy theory and the Bayes theorem have been recognized more and more in high quality data analysis.Obscure description of hypotheses accurately depict people’s proposition treating hypotheses testing,so under the view of the fuzzy theory to study the corresponding hypothesis test problem has become the concern of many scholars in recent years.Based on this,this paper chooses the fuzzy hypothesis test as the research object,and studies the determination of the corresponding membership function and the Bayes test and inference method.Determination of membership function in the existing studies are often identified as empirical problems,so firstly we take piecewise linear membership function as the research object,based on fuzzy P values the Bootstrap method of the classical statistical theory,we combine the subjective cognition with sample information to determine membership function with quantitative way.Secondly,base on cosine type membership function which is more fit hypothesis connotation,we present the fuzzy hypothesis test under the theory of Bayes theorem in the Bootstrap method.On the one hand,by the Bootstrap density function to sample on the sample information,in order to improve the hypothesis can distinguish between sex,reduce the vagueness of the decision;On the other hand,in the determination of membership function,the parameters of the Bootstrap method important characteristic value,the standard deviation of the replication value as a parameter to improve the reliability of the decision.Lastly the study is based on probability distribution of membership functions,by assuming the information added to the statistical inference process,make the fuzzy hypothesis test under the Bayes theory conforms to set it as a way of statistical inference.