摘要： 相似文本的快速搜索是大规模文本处理需要解决的基本问题。从两方面改进了Udi的相似文本搜索方法,通过Hash把集合映射成ID,从而得到更快的集合比较算法,重新定义了相似关系,能够减少误判,同时对有固定格式的文本也有更好的效果。

关键词: 大规模文本处理, 相似文本搜索, 复制检测

Abstract: Searching similar texts is a fundamental problem for many large scale text processing tasks. Udi's algorithm for searching similar texts is improved in two ways. By mapping each set to an [D, a faster algorithm to compare sets is obtained. And the relation of similar to is redefined in order to both reduce false decision and improve the performance for those texts with fixed format.

中图分类号: 

• TP391.43