Association-Rules-Based Recommender System for Personalization in Adaptive Web-Based Applications

Personalization systems based upon users' surfing behavior analysis imply three phases: data collection, pattern discovery and recommendation. Due to the dimension of log files and high processing time, the first two phases are being achieved offline, in a batch process. In this article, we propose WRS, an architecture for adaptive web applications. Within the framework, usage data is being implicitly achieved by data collection submodule. This allows for the extraction of usage data, online and in real time, by using a proactive approach. For the pattern discovery, we efficiently used association rule mining among both frequent and infrequent items. This is due to the fact that the pattern discovery module transactionally processes users' sessions and uses incremental storage of rules. Finally, we will show that Wise Recommender System (WRS) can be easily implemented within any web application, thanks to the efficient integration of the three phases into an online transactional process.