摘要
提取合适的正例文本是生成有效过滤模板的重要基础。通常采用的人工评测方法的显著缺点是无法准确判断文本内容并选择切实相关的文本,所以本文在频繁项集挖掘的基础上提出了一种判断给定文本是否相关并提取合适正例文本的方法。实验结果证实了该方法的有效性。
It is the important groundwork for constructing an effective filtering profile to extract right positive - example text. The method of Manual evaluating used usually can not distinguish the contents of texts and vote in correlative texts well, so this paper brings forward a method for distinguishing the contents of texts and voting in correlative texts well based on frequent itemset mining. The result of experiments proves its effect.
出处
《信息技术与信息化》
2008年第3期41-43,共3页
Information Technology and Informatization