摘要
分析了普遍搜索引擎不能为用户提供具有个性化服务的原因,提出了基于页面内容与链接的页面价值快速算法,给出了算法的基本思想及对应的模型,并通过计算以转移概率矩阵为系数方程的特征值得到页面的价值。结果表明,新的模型能够以较少的计算量达到类似TFIDF算法的查全率。
The reason of the general search engine cannot provide the personalized information retrieval service had been discussed. This paper provided the quick algorithm of page value computing, the basic idea and the model of the algorithm. The page value could be computed as the eigenvalue of the transition probability matrix. The experiment had been done based on the WT10g data set. The result showed that the new model could reach the maximize recall just as the TFIDF with the less computing work.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第8期139-141,共3页
Microelectronics & Computer
基金
陕西省自然科学基金项目(2005F08)
陕西省教育厅专项基金项目(06JK300)
关键词
个性化模型
类关键词
转移概率矩阵
页面价值
personalized information retrieval
class keywords
link matrix
page value
