Symmetric deviations and distance measures

Abstract. In this paper we characterize measurable information measures depending upon two probability distributions in a unified manner to get most of the existing information measures. Moreover it turns out that our characterization contains new, unexpected information measures.