A Flexible System of Precedence Coefficient Aggregation and Consensus Measurement

The paper presents a practical method for a software meant for the in-session processing of precedence data obtained from m judges and concerning n items. Precedence data of a judge can be given in a variety of ways — from a unique ordering to 1/2;-n (n-1) fuzzy precedence indices. First, the principles of processing are presented, then problems and suggested solutions. The latter refer to two notions: “expert agreement” as to the final precedence coefficients on the one hand and “resulting ordering” of the items on the other.