GROUPER: A DYNAMIC CLUSTERING INTERFACE TO WEB SEARCH RESULTS

1

• GROUPER A DYNAMIC CLUSTERING INTERFACE TO WEB
  SEARCH RESULTS
• by
• Oren Zamir and Oren Etzioni
• Presented by
• Duygu Sarikaya,Ahsen Yergök,Dilek Demirbas

2
OUTLINE

• Motivation
• Problem Definition
• Proposed Solution Goals
• How it Works
• User Interface
• Easy to Browse Clusters
• Design for Speed
• Empirical Evolution
• Conclusion

3
PROBLEM DEFINITON

• Search engine results are not easy to browse

4
PROPOSED SOLUTION

• GROUPER
• An Interface
• Dynamic Clustering of Search Results

5
GOALS

• Coherent Clusters
• Efficiently Browsable
• Speed

6
User Interface

• Groupers query interface

7
A Query Result
8
Refine Query Based On This Cluster
9
MAKING THE CLUSTERS EASY TO BROWSE
Identifying Redundant Phrases

• Word Overlap
• Sub- and Super- Strings
• Most-General Phrase with Low Coverage

10
DESIGN FOR SPEED
11
EMPIRICAL EVALUATION OF GROUPER
12
Comparison to a Ranked List Display

• Compared with HuskySearch, but not TREC
• Number of documents followed
• Time spent
• Click distance

13
TIME SPENT

• The average amount of time spent on each document

14
CONCLUSION

• Grouper
• Comparison to the logs of HuskySearch
• Document clustering system on the Web
• Grouper-II