Sigmoidally Constrained Maximum Likelihood Estimation in Quantal Bioassay

From MaRDI portal





The probability that an arbitrary subject responds in a specified way to a dose x is denoted by \(\pi\) (x). A nondecreasing dose-response curve \(\pi\) is sigmoid if there exists a point of inflection. Under the condition that \(\pi\) is sigmoid, maximum likelihood estimates and approximate and exact upper confidence bounds for \(\pi\) (x), especially for small \(x_ i\) are discussed. These depend on an inequality that holds when \(\pi\) is convex on \([0,x_ 1]\). A test of concavity against the alternative (sigmoidicity) is discussed. Rejecting concavity requires evidence that \(\pi\) is convex on \([0,x_ 1]\).











This page was built for publication: Sigmoidally Constrained Maximum Likelihood Estimation in Quantal Bioassay

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3347165)