正態(tài)模型刻度參數(shù)的經(jīng)驗(yàn)貝葉斯估計(jì)(1)(留存版)
【正文】 The theoretical Bayes evaluation was indicated by using X marginal distribute density function and its first order derivative . Using the past sample value (x1 x2….xn) and present value x , bined with kernel estimation method to construct relevant function to replace the function in Bayes evaluation , the parameter by empirical Bayes was obtained . It was proved that the proposed estimator was an asymptotical optimal EB estimator. Key words: general normal mode parameter empirical Bayes kernel estimation asymptotical optimality 一、問(wèn)題的提出 Bayes統(tǒng)計(jì)推斷原則：對(duì)參數(shù)所作任何推斷必須基于且只能基于的后驗(yàn)分布，即后驗(yàn)密度函數(shù)族，它依賴于的先驗(yàn)分布，而先驗(yàn)分布往往很難確定，1955年R0bbins提出了經(jīng)驗(yàn)貝葉斯（EB）方法．自這種方法提出以來(lái)，EB估
研究報(bào)告相關(guān)推薦
鄂ICP備17016276號(hào)-1