Research on Association Rules Based Group Ranking Model with Fuzzy Preference Relation Structure

In this paper, we propose a method that takes full advantage of the characteristic that association rules can reflect the logic relationships between items in the frequent itemsets for ranking the numbered alternatives. In group ranking, the decision data are usually pairwise comparisons based on fuzzy preference relations. The concordance and discordance factors are pivotal to the quality of the solutions. And fuzzy preference relations usually can not hold the consensus in the real situation. In our method, all the preference information that is obtained by abstracting association rules from the frequent itemsets of pairwise comparisons is considered as contribution to the consistent measures of the solutions, which shows the majority rule of the collective based on statistic and makes the solutions obtained more objectively true. An example is given to show the availability, which is compared with the result of the method using the ordered weighted averaging (OWA) operator