Abstract: Text knowledge mining plays a very important role in data mining.In the paper，we mainly discuss how to extract concept lattices from different texts without using conceptual scaling and to analyze the structural connections among concept lattices.Our method consists of two parts：one is to transform the objects described in texts into many-valued contexts，and the other is to analyze some operations such as addition，deletion and product and the connections among concept lattices such as order-embedding map.We obtain some important propositions，which can be used by knowledge engineers to analyze text knowledge and to further extract associate rules from concept lattices.

Key words: text knowledge, data mining, many-valued contexts, formal concepts, concept lattices, structural connections

摘要： 文本知识挖掘是数据挖掘中一个很重要的研究领域。论文主要讨论如何在不使用概念换算方法下从文本知识中抽取概念格以及分析概念格之间的结构关联。该方法有两部分构成：一是将文本中所描述的对象转化为多值上下文；二是分析多值上下文之间的各种操作以及相应概念格之间的关联。重点分析了多值上下文的增加、删除和乘积等操作以及相应概念格之间的序嵌入映射，得到了一些重要命题。知识工程师可以利用这些命题进行文本知识分析以及从概念格上进一步抽取关联规则。

关键词: 文本知识, 数据挖掘, 多值上下文, 形式概念, 概念格, 结构联通