论文部分内容阅读
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。
On the basis of fuzzy clustering, the distribution characteristics of fuzzy transfer relations between multidimensional rules are analyzed and a dimension-by-dimension mining algorithm (MMFCRA) with multi-level fuzzy rules is proposed. The algorithm generates a multidimensional fuzzy rule set by gradually reducing the recursive computation of the candidate set, thereby establishing a multi-dimensional and multi-level fuzzy rule mining model, which can effectively reduce the number of query computations without reducing the number of dimensions, It reflects the influence of the interval dimension combinations of different dimension indexes, and at the same time, guarantees the convergence of the dimension-by-dimension mining algorithm and avoids the loss of knowledge.