Abstract: Knowledge acquisition is one of the necessary and important steps in text mining, and it has been a common way to acquire knowledge from texts. However, different texts knowledge resources may lead to knowledge inconsistency. In order to acquire consistency knowledge, this paper analyzes all kinds of inconsistency in texts, and provides corresponding inconsistency detection and revision algorithms. This step is very important for knowledge service. The experimental results show that the method is effective and feasible.

Key words: text mining, frame representation, knowledge inconsistency, knowledge validation

摘要： 知识获取是文本挖掘中必要且重要的步骤之一，从文本中获取知识已成为了一种常用途径，然而不同文本知识源可能引起知识的不一性问题。为了保证所获取知识的一致性和精确性，分析了文本知识的不一致性类型，并提出了基于框架的不一致性检测和修正算法，为下一步的知识服务奠定了基础。实验结果表明所提出的不一致性处理方法是可行和有效的。

关键词: 文本挖掘, 框架表示法, 知识不一致, 知识修正