Search

Abandonement rates of patents in five European countries are studied. The time distribution of British patents can be described by a truncated Gauss' distribution, of Belgian patents by a Poisson's distribution. Appropriate models derived on the base of the octogonal and cubical linear vector space norms are discussed. [ABSTRACT FROM AUTHOR]

A linear correlation exists between the Lotka frequency and Zipf rank distribution functions. Relatively good correlation coefficients were found, but slope constants are not consistent with theory. They show that information distributions are not homogeneous and cannot be completely described by two parameter functions. [ABSTRACT FROM AUTHOR]

Haitun's conjectures about specific properties of information distributions are questioned. It is shown on linguistic material that the opening of a closed counting scale changes the skewness of distribution in a way which contradicts Haitun's suggestions. Negative power distributions are time dependent and can be explained as density measures of fractal clusters and are not specific to the social sciences. [ABSTRACT FROM AUTHOR]

Information spectra are defined as intervals between equivalent information events. Their relations to negative binomial and negative polynomial distributions and urn models are explained. Basic properties of empirical information spectra from patent literature are shown and discussed in connection with Haitun's views on Z type information distributions, Sichel's GIGP model and Trofimenko's study on formation and decay of author groups. [ABSTRACT FROM AUTHOR]