Abstract:

For the difficult determining of the membership function and the hard border of interval division when the fuzzy association rules are mining,association rules mining algorithm based on fuzzy relation is proposed.After the relation grade numbers and the similarity of neighbor grades are determined,the linguistic expressions （the properties of transactions）are mapped to each grade of the label set and then get the grade ratings as the weights.Based on these weights,the fuzzy relation support and confidence are defined.The detailed steps of the algorithm are elaborated in this paper.At the end of this article the application of this proposed algorithm about mining association rules in service trust area is described.

Key words: linguistic value, fuzzy relation, association rule, service trust area