THE SOCIAL FUTURE OF WEB SEARCH: MODELING, EXPLOITING, AND SEARCHING COLLABORATIVELY GENERATED CONTENT
Eugene Agichtein, Evgeniy Gabrilovich, Hongyuan Zha · 2009
Social, or collaboratively generated content (CGC) is transforming how we seek and find information online: it is now a prominent part of the web information ecosystem, and a powerful platform for information seeking. The resulting archives of both the content and the context of the interactions contain valuable information that is often not available elsewhere, and can be helpful for the development of novel ranking algorithms, and natural language processing, text mining, and information retrieval techniques. We review machine learning techniques for modeling CGC, focusing on tasks such as learning to estimate content quality, relevance, and searcher intent and satisfaction with the retrieved results. We describe how this information can be incorporated into learning-based ranking methods for searching social media, and how CGC could be used to improve performance on key text mining and search tasks. 1