摘要
本文针对基于关联规则的文本过滤器设计做了如下探索:(1)针对中文网络语言的特点,引入n-Gram方法提取文本的特征;(2)提出边界样本的概念;(3)在设计基于关联规则的文本过滤器时,引进了否定选择算法,采用否定选择算法对过滤器的检测器集合进行自体耐受,最终建立高准确率的文本过滤器。实验表明,经过自体耐受的过滤器可以有效地提高过滤准确率。
As for the text filter design based on association rules, the paper makes the following efforts： （1）As for the charateristics of the Chinese web language, we introduce the n-Gram method to extract text features; （2）We propose the concept of edge sample; （3）When designing the text filters based on association rules, we introduce a negative-selection algorithm to make the filters＇ detector set tolerant, and finally build a high-precision text filter. Experiments show that the filters after proper toleration can effectively increase the precision of filtering.
出处
《计算机工程与科学》
CSCD
2008年第8期61-64,共4页
Computer Engineering & Science
