
APPENDIX 1
Informational Entropy
ABSTRACT
Objects are described by digital vectors of properties. A hierarchical order
for digits is considered. The probabilities are replaced by similarities
between vectors. Similarity matrices may be stabilized by max-min
procedures. Classifi cation is based on similarity and distance evaluation.
Informational entropy and informational distance are defi ned.
Keywords: classifi cation, distance, entropy, information, similarity, vector
of properties
The initial step in quantifying the concept of similarity for objects, as for
instance the chemical species in a mixture, is to list the most important
structural elements ...