Correction of Renyi´s `gain of information´

A number of measures have been proposed for the divergence (discrepancy, “distance”) between two probability distributions. The first such generalized (parameterized) measure of divergence, or `gain of information´, was proposed by Renyi (1961, 1970). However, Renyi´s measure is quite restrictive in terms of certain conditions that need to be met. The purpose of this paper is to suggest some corrections, or modifications, to Renyi´s measure to make it less restrictive and potentially more useful