Integration Challenges for a Web-based Personalized Query Suggestions System in Information Retrieval

We analyze in this paper the design architecture of a web-based personalized query suggestion system usable in information retrieval contexts. Our system comes in the form of a transparent proxy for the Google search engine (although it can work with other search engines too, with minimal code changes) that is able to personalize the list of query suggestions provided by Google, based on the recent user web browsing history. The most important design choices are presented with a special focus on the integration with the Google search engine. The integration analysis is directed by the well known application integration criteria. We have made our system transparent to the user when searching information on Google and we have shown through performance tests that our system has very small computational costs for the user experience.