Abstract:
In the same domain, different Web databases have overlap with each other. In order to estimate the overlapping rate in Web database, this paper proposes an attribute high-frequency words based approach for estimating the overlapping rate of Deep Web database. Starting with the Chinese frequency words, the attribute frequency words can be induced iterative, and using its query results calculate the conditional probability of overlapping part, then estimate the overlapping rate. Experiments prove its bias and volatility are less than 5.
Key words:
overlapping rate,
Deep Web,
database,
estimation
摘要: 在同一领域中,不同Web数据库之间有重叠部分,为估计重叠部分在Web数据库中所占比例,提出一种基于属性高频字的重叠率估计方法。以中文高频字作为起点,迭代诱导属性高频字。通过属性高频字的查询结果计算重叠部分出现的条件概率,进而估计重叠率。实验证明,该方法的偏差度和波动度都小于5。
关键词:
重叠率,
深网,
数据库,
估计
CLC Number:
MIAO Zhong-yi; ZHAO Peng-peng; HU Peng-yu; CUI Zhi-ming. Estimation for Overlapping Rate of Deep Web Databases Based on Attribute High-frequency Words[J]. Computer Engineering, 2009, 35(13): 28-30.
苗忠义;赵朋朋;胡鹏昱;崔志明. 基于属性高频字的深网数据库重叠率估计[J]. 计算机工程, 2009, 35(13): 28-30.