%0 Conference Paper %B IADIS European Conference on Data Mining 2009, Algavre, Portugal %D 2009 %T Date - based dynamic caching mechanism %A Christos Bouras %A Vassilis Poulopoulos %A Panagiotis Silintziris %X News portals based on the RSS protocol are becoming nowadays one of the dominant ways that Internet users follow in order to locate the information they are looking for. Search engines, which operate at the back-end of a big portion of these web sites, receive millions of queries per day on any and every walk of web life. While these queries are submitted by thousands of unrelated users, studies have shown that small sets of popular queries account for a significant fraction of the query stream. A second truth has to do with the high frequency that a particular user tends to submit the same or highly relative search requests to the engine. By combining these facts, in this paper, we design and analyze the caching algorithm deployed in our personalized RSS portal, a web-based mechanism for the retrieval, processing and presentation in a personalized view of articles and RSS feeds collected from major Internet news portals. By using moderate amounts of memory and little computational overhead, we achieve to cache query results in the server, both for personalized and non-personalized user searches. Our caching algorithm operates not in a stand-alone manner but it co-operates and binds with the rest of the modules of our Portal in order to accomplish maximum integration with the system. %B IADIS European Conference on Data Mining 2009, Algavre, Portugal %P 67-74 %8 June 18 - 20 %G eng