An approach to the information analysis is considered for the case when the information is presented by words of finite length over a finite alphabet. A method of generating a measure of symbolic diverseness of words based on peak characteristics of a shift entropy function is proposed. The shift entropy function is formally defined using a unit translation operator and the entropy of discrete distributions. A model example is presented together with some results of application of the proposed measure in the clustering of families of plants using the analysis of genome of their representatives.