元组统计相似性知识的提取与应用

A METHOD OF EXTRACTING OBJECTS' SIMILARITY KNOWLEDGE FROM DATABASES

在实际应用数据库中，结合少量领域专家的定性知识，自动、高效地提取并利用元组间相似性知识，能够建立更健壮的知识系统．文中讨论了在复杂背景条件下，基于统计的定性数据相似性比较问题，提出了属性取值相似性知识动态提取和利用的新方法，介绍了该方法在公安刑侦案例分析上的实际应用．

Extracting and making use of the similarity knowledge automatically and efficiently from large databases can help to build more robust knowledge systems. This paper is concerned with the similarity comparisons of catalog data based on statistics, and a new method is proposed to extract and use the attribute similarity knowledge dynamically. The method has been used in case analyses in Beijing Public Security Bureau.