Path complement is an important phase in Web log data preprocessing. Most of the existing path complement technology is implemented based on static site structure. Due to the extensive use of personalized recommendation technology, the site structure is converted from static structure to dynamic structure. Considering that existing path complement algorithms can not solve the problem of page lost in user access path in dynamic site structure, this article presents the concept and the construction method of dynamic site structure as well as the graph structure storing policy of the site structure. Based on these, this article presents a path complementing algorithm based on page sorts (PCBPS) in dynamic site structure. Experiments prove that this algorithm can recover lost pages in user access path accurately and well improve the accuracy of path complement.%路径补全是Web日志数据预处理的重要阶段,目前的路径补全技术大多基于静态网站结构实施.个性化推荐技术的广泛应用,使站点结构由静态结构转变为动态结构.针对目前各种路径补全算法无法解决动态站点结构下用户访问路径中页面缺失的问题,提出动态站点结构的概念、构造方法及站点结构的图结构存储策略.在此基础上,提出一种在动态站点结构下的基于页面类型的用户访问路径补全算法PCBPS(Path Complement Based on Page Sort).实验证明在动态站点结构下,这种方法能较准确地恢复用户访问路径中的缺失页面,较好地提高了路径补全的准确率.
展开▼
