An Analysis Study of Vertical Selection Task in Aggregated Search
Sanae Achsas, El Habib Nfaoui · Procedia Computer Science · 2019
Aggregated Search (AS) paradigm is generally decomposed into two sub-tasks, the first one is Vertical Selection that focuses on predicting which verticals, if any, are relevant given a user query. The second is Results Presentation, that aim at deciding where to present the different results of the selected verticals. In this study, we have focused on vertical selection sub-task, which is considered as the most important task in the Aggregated Search process, and that because the task of presenting the relevant results can only be done after the selection of the relevant verticals. But despite its interesting role, there is no huge literature regarding vertical selection, as it is difficult to assess the open research problems and the real advancement in this research field. The purpose of this article is to assemble and analyses the different efforts made in Vertical Selection to the date. This includes different sources of evidences used to make relevant predictions, as well as various vertical selection approaches and evaluation methodologies exploited in the literature. Research in this field has shown an evolution from 2008 to 2013, in the next three years (2013-2016), this area didn’t receive significant attention from researchers. In the last two years, the vertical selection research has been renewed. This analysis study aims at better understanding and clarifying the advancement of this task which can bring new ideas to improve its performance in future works.