SVM and AdaBoost Based Ranking Model Adaptation for Domain Specific Search

Abstract: The invent of Web 2.0 has enabled building of state of the art web applications including sophisticated search engines. The web applications are capable of providing domain specific search which enables end users to gain access to required data quickly. However, the results might be bulky and may not be relevant to the user-intended results. In other words users have to spend some time browsing results for finding required information. To overcome this problem many ranking algorithms came into existence. The ranking algorithms help users to find required results quickly. But the ranking models in the existing work were built based on broad-based ranking which is not useful for other domains. Recently Geng et al. proposed a ranking model adaption framework which can adapt to various domain specific searches. They used SVM for building ranking model. In this paper we built a prototype application that demonstrates ranking model adaption using a novel ranking model meant for ranking the search results besides adapting to new domains. The experimental results revealed that the proposed application is useful in searching data across the domains.

Keywords: Ranking model, domain specific search, SVM, model adaptation