An OLAP-based Scalable Web Access Analysis Engine

摘要: Collecting and mining web lag records (WLRs) from e-commerce sites has become increasingly important for targeted marketing, promotions, traffic analysis. In this paper, we describe a scalable data werehousing OLAP-based engine analyzing WLRs. We have to address several scalability performance challenges in developing such framework. Because an active site may generate hundreds of millions WLRs daily, deal with huge volumes flow rates. To support fine-grained analysis, e.g., individual users' access profiles, end up huge, sparse cubes defined over very large-sized dimensions (there be hunderds thousands visitors the tens pages). While OLAP servers store quite efficiently, rolling large cube can take prohibitively long. applied non-traditional approaches problem, which allow us speed WLR analysis by 3 orders magnitude. Our framework multilevel multidimensional pattern extraction, feature ranking, addition typical operations, supports operations as extended association rules.