Pagerank是评定网页等级的一种比较流行、比较重要的方法,但是随着网页数量指数增加,爬行所有的有用的网页变得越来越困难,结果是,网页的等级值的计算仅仅依靠整个网络中的一个子集,这样产生了不正确结果.因为存在不完整的信息(悬挂页面),这个情况是存在于计算当中的,为了克服这个不完整性,文章提出了pagerank算法的一种改进-页面预测分类法,并对此算法进行效率分析.这可以将不同类别的悬挂页面进行特殊的分析以至于整个链接结构能够比较准确的预测,而且,实验结果显示:与先前的pageramk算法相比这个算法达到了令人比较满意的结果.%The Pagerank is a popular, more important method of evaluation Pagerank. But as the number of web index increased, crawling all useful page becomes more and more difficult. Thus, The calculated value of the page rank the whole network just rely on a small subset, thus produced incorrect results. Because of the existence of incomplete information(Pensile page), this situation is exist in calculation. In order to overcome this, we proposed an improved algorithm of pagerank,called prediction classification, and the efficiency of this algorithrn. This can be different types of pensile page that the link analysis of the special structure can be accurately predicted, and that, experimental results show, compared with the previous pagerank algorithm in this algorithm is achieved satisfactory results.
展开▼
