Abstract: In this paper，based on G poset，M poset and GM poset in FCA，we propose several basic concepts such as object upper neighbor，object lower neighbor and maximal lower boundary of object，and some related theorems.Then by means of the concepts and the theorems，an algorithm for constructing line diagrams of the posets is presented.As an application for data mining，not constructing a concept lattice，we show an algorithm to directly extract implications from the obtained line diagrams and an example illustrates the effectiveness of the algorithm.

Key words: formal concept analysis, poset, concept lattice, concept

摘要： 对形式背景上的G偏序集、M偏序集和GM偏序集的偏序关系作了进一步讨论，提出了对象上邻、对象下邻、属性上邻、属性下邻、对象上界、对象下界、属性上界、属性下界和极大对象下界等概念，并根据上下邻之间的关系给出G偏序图、M偏序图和GM偏序图的构建算法，最后给出了不需要建格而可直接在GM偏序图上进行规则提取的算法，并通过实例验证了算法的有效性。

关键词: 形式概念分析, 偏序集, 概念格, 概念