Abstract:  Formal concept analysis is a powerful tool for data analysis and extracting rules from formal context. The attribute topology, as a novel representation of formal context, describes the association between attributes visually. Using the attribute topology theory to compute formal concepts and concept lattices is more convenient and intuitive. By studying the existing algorithms and processes based on attribute topology, the specificity and level-limitations are analyzed in this paper. By combining the concept of degrees in graph theory, a new way, attribute degree, to measure attributes is proposed. And based on attribute degree, a novel attributes-sorting algorithm is proposed. This attribute degree-based sorting algorithm is more flexible, eliminating level-limitations of attributes sorting. Parent attributes can be easily found, and new algorithms of attribute topology can be explored.

Key words: attribute topology, attributes-sorting, attribute degree, formal concept analysis

摘要： 形式概念分析是一种从形式背景进行数据分析和规则提取的强有力工具。属性拓扑作为一种新型的形式背景表示方法，直观地描述了属性之间的关联。利用属性拓扑可以更方便直观地计算形式概念和概念格。经过对现有属性拓扑的算法与流程的研究，分析了现有属性排序算法的特异性和层次局限性，通过结合度的概念，提出了一种属性衡量的新方式——属性度，并提出了基于属性度的属性排序算法。这种排序算法得到的结果更加灵活，消除了属性排序的层次局限性，对父属性的查找有明显的优势，为基于属性拓扑中的属性排序方法提供了指引方向。

关键词: 属性拓扑, 属性排序, 属性度, 形式概念分析