Effective extraction of data association rules can provide a reliable basis for classification of stellar spectra. The concept of stellar spectrum weighted itemsets and stellar spectrum weighted association rules are introduced, and the weight of a single property in the stellar spectrum is determined by information entropy. On that basis, a method is presented to mine the association rules of a stellar spectrum based on the weighted frequent pattern tree. Important properties of the spectral line are highlighted using this method. At the same time, the waveform of the whole spectrum is taken into account. The experimental results show that the data association rules of a stellar spectrum mined with this method are consistent with the main features of stellar spectral types.
Jiang-Hui CaiXu-Jun ZhaoShi-Wei SunJi-Fu ZhangHai-Feng Yang
约束概念格是概念格的特化结构,构造时具有较低的时空复杂度,能从中快速提取比较丰富的信息和知识.为了提取分类规则,在充分分析约束概念格结点外延与数据集等价划分之间关系的前提下,引入了分类支持度和记录支持度的概念,提出了一种面向约束概念格的分类规则提取算法(Classification Rule Acquisition Algorithm based on Constrained Concept Lattice,CRACCL),并采用UCI数据集作为实验集,验证了本算法能够提取更加实用和准确的分类规则.
