Robert Hafner

An Improvement of the Harman-Fukuda-Method for the Minres Solution in Factor Analysis.

The method proposed by Harman and Fukuda to treat the so-called Heywood case in the minres method in factor analysis i.e., the case where the resulting communalities are greater than one, involves the frequent solution of eigenvalue problems. A simple method to treat this problem requiring less computing time and enjoying higher numerical stability is described in this paper.

Simple construction of least favourable pairs of distributions and of robust tests for prokhorov-neighbourhoods

The construction of least favourable pairs of distributions for testing problems, where one hypothesis is simple and the other composite of the Prokhorov-neighbourhood-type has been given by Osterbeicheb in [1] for the case where the sample space is . In this paper the case where both hypotheses are composite of the Prokhorov-neighour-hood-type is treated and a direct construction of a least favourable pair of distribution is given. The applicability of the method, especially the construction of robust tests, is demonstrated for an example.

Construction of minimax-tests for bounded families of probability-densities

SummaryA new method for the construction of least favourable pairs of densities and of minimaxtests is given for the compound test-problem

Grundbegriffe der Wahrscheinlichkeitsrechnung

H_0 :\underset{\raise0.3em\hbox{

Maßzahlen für mehrdimensionale Verteilungen

\smash{\scriptscriptstyle-}

Der Chi-Quadrat-Test

}}{g} \leqslant f \leqslant \bar g

Gesetze der großen Zahlen

against