【摘要】 提出一种基于广义后缀树的概念生成算法(generalized suffix tree based concept generation algorithm,GSTCG)，将背景中所有对象的属性序列及其后缀建立为一棵广义后缀树，并根据广义后缀树产生候选概念；其次，合并具有相同对象集合的候选概念，再根据规则对候选概念进行扩展；最后，删除冗余的候选概念后得到全部形式概念。在两类不同参数人工数据集上的实验结果表明，GSTCG算法与NextClosure算法在所有背景上得到的概念数量一致，且前者具有更优的时间性能。更多还原

【Abstract】 A generalized suffix tree based concept generation algorithm(GSTCG) is proposed. Firstly, attribute sequences of all objects in the background and their suffixes are established as a generalized suffix tree, and candidate concepts are generated according to the generalized suffix tree.Secondly, candidate concepts with the same object set are combined, and then candidate concepts are extended according to rules.Finally, all formal concepts are obtained after redundant candidate concepts are deleted. Experimental results on two kinds of artificial data sets with different parameters show that GSTCG algorithm and NextClosure algorithm get the same number of concepts in all backgrounds, and the former algorithm has better time performance.更多还原