An Efficient Bayesian Optimal Design for Logistic Model

An Efficient Bayesian Optimal Design for Logistic Model

Consider the Bayesian optimal designs with many support points such that arise the problem of having small number of observations at each design point. Under such a scenario the asymptotic property of using Fisher information matrix for covariance matrix of ML estimators might be doubtful. We used Bhatteharyya matrix as an approximation for covariance matrix to proposed a modified Bayesian D-optimal criterion in construction a more efficient optimal design for logistic model. It is shown that the resulted optimal designs are more efficient than the previous one. obtained under the same conditions.