一个多层次模糊规则的逐维挖掘算法

在模糊聚类的基础上,分析多维规则之间模糊传递关系的分布特性,提出一个多层次模糊规则的逐维挖掘算法(MMFCRA).该算法通过逐步精简候选集的递推计算来生成多维模糊规则集,从而建立多维多层次的模糊规则挖掘模型,能够在不减少维数的情况下,既有效地降低了查询计算的次数,又反映了不同维数指标区间组合的影响情况,同时也保证了逐维挖掘算法的收敛性,避免了知识的遗失.

Dimension Mining Algorithm of Multidimensional and Multilevel Fuzzy Rule

By analyzing the distribution features of fuzzy transitivity in multidimensional data on the basis of fuzzy clustering,this paper proposes a dimension mining algorithm of multidimensional fuzzy rule(MMFCRA).The algorithm builds a collection of multidimensional fuzzy rules through recursive calculation of reducing candidate intemsets,and subsequently establishes a multidimensional mining model.Without reducing the dimension,this algorithm not only efficiently cuts down the number of inquiry computation,but also reflects the influence of combination of different dimension indexes,and guarantees the convergence of MMFCRA as well,avoiding the loss of knowledge.