A Study of an Improved Data Pre-processing Method Based on Web Usage Mining
PU Hui-zhon · 2014
Based on the various data collecting methods, this paper analyses their respective limitations. First, the server is determined for Web usage mining. Then the characteristics of the server data sources are analyzed. For the data sources, a data pre-processing program and an improved method are put forward. The key pre-processing steps such as data cleaning, user identification, session identification, etc. are proposed, which provides accurate target data for the data mining process.