Abstract: Each visitor's shopping session of the E-Business Web site is recorded in the Web server log files. Analyzing the log files and exploring the strong regularities in the commodities of the shopping cart, can provide the recommended goods for Web users, and improve the performance of the Web service. In order to improve the privacy preservation of the original visitor's shopping information and mining result, an effective method for privacy preserving association rule mining was presented. First, a new data preprocessing approach, Fake Column' s Randomized Response with Column Replacement (FCRRCR) was proposed to transform and hide the original data. Then, an effective privacy preserving association rule mining algorithm based on bit AND operation was presented. As shown in the experimental results, the algorithm can achieve significant improvements in terms of privacy, accuracy, efficiency and applicability.

Key words: Web logs, Privacy preservation, Association rule, Randomized response