【摘要】 针对标准Fish算法的链接相关度区分度低、搜索宽度不灵活、主题描述不准确的缺点,提出一种改进的算法。在计算链接相关度时,引入网页分块的思想以及增量反馈和自适应机制。在计算搜索宽度时,提出一个启发性规则来动态改变其大小。实验结果表明,改进算法的查准率和查全率分别提高了118%和24.16%。更多还原

【Abstract】 Since the defect of the standard Fish algorithm that the differentiation of link-relevance is low and the searching width lacks flexibility and topic description is not accurate,this paper proposes an improved algorithm. When computing the link relevance,it introduces the webpage-partition thought and the mechanism of increment feedback and self-adapting. When computing the searching width,it proposes an enlightening rule to change its value dynamically. Experimental results show that the improved algorithm increases the precision ratio and recall ratio by 118% and 24.16% respectively.更多还原