scientific article; zbMATH DE number 4003284

From MaRDI portal
Publication:4728043





[For the entire collection see Zbl 0615.00010.] This paper presents results from our research program on the development of statistical methods based on quantile-information-function estimation statistical inference. Its major goal is to define and to apply the comparison quantile function D(u;F,G) of two distribution functions F and G. We outline how standard methods of fitting a parametric model F(y,\(\theta)\) to one or two random univariate samples can be developed in terms of the empirical quantile function \(Q^{\sim}(u)\) by suitably defining the sample comparison quantile function \(D^{\sim}(u;\theta)=D(u;F^{\sim},F(\cdot,\theta)\) and the sample comparison density function \(d^{\sim}(u;\theta)=D^{\sim}{}'(u;\theta)\) for both continuous and discrete data. In the continuous case we define \(D^{\sim}(u;\theta)=F(Q^{\sim}(u);\theta)\). In the discrete case we define first \(d^{\sim}(u;\theta)\) and then define \(D^{\sim}(u;\theta)\) to be its integral. Other concepts defined include: continuous versions of discrete quantile functions, identification quantile function, information distributions, and scientific statistical science. Emphasis is placed on developing a notation which applies to both discrete and continuous data analysis.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4728043)