Search Snippet Evaluation at Yandex: Lessons Learned and Future Directions

This papers surveys different approaches to evaluation of web search summaries and describes experiments conducted at Yandex. We hypothesize that the complex task of snippet evaluation is best solved with a range of different methods. Automation of evaluation based on available manual assessments and clickthrough analysis is a promising direction.

• Type: Journal
• Year: 2011
• Where: CLEF
• Authors: Denis Savenkov, Pavel Braslavski, Mikhail Lebedev
• Keywords: Automation, CLEF 2011, Information Technology, Promising Direction, Search Summaries