A framework for personalized and collaborative clustering of search results
Proceedings of the 20th ACM international conference on Information and knowledge management
| Hi-index | 0.00 |
With a Wiki-like search interface, users can edit ranks of search results and share the edits with the rest of the world. This is an effective way of personalization, as well as a practice of mass collaboration that allows users to vote for ranking and improve search performance. Currently, there are several ongoing experimentation efforts from the industry, e.g., SearchWiki by Google and U Rank by Microsoft. Beyond that, there is little published research on this new search paradigm. In this paper, we make an effort to establish a framework for rank editing and sharing in the context of web search, where we identify fundamental issues and propose principled solutions. Comparing to existing systems, for rank editing, our framework allows users to specify not only relative, but also absolute preferences. For edit sharing, our framework provides enhanced flexibility, allowing users to select arbitrarily aggregated views. In addition, edits can be shared among similar queries. We present a prototype system Rants, that implements the framework and provides search services through the Google web search API.