ABSTRACT
Classifi cation has played an important role in Archaeology. In fact, one of the main aims of this discipline is to determine and establish the laws that order the diversity of material culture. Archaeologists have made use of qualitative values and, since the 1960s, of quantitative techniques in the so-called Numerical Taxonomy, with analytic methods mainly of classifi cation, and discriminant analyses playing a major role. These applications, fi rstly developed by R.A. Fisher in 1936, have been of great importance because they search for lineal combinations of variables to characterize or distinguish observation in two or more groups. Also, this is because their use can establish the probability that one observation will belong to a previously defi ned group. Therefore, two of the main characteristics of discriminant analyses are: 1) their ability to maximize the differences between groups, and 2) their predictive character, i.e., the ability to assign individuals to groups (predictive discriminant analyses).
