A method for improving the biostatistical calculus of paternity based on the entropy of DNA polimarker allele distribution