Reputation Metrics in Opinion Space
Ephrat Bitton, Siamak Faridani, David Wong and Ken Goldberg
Berkeley Center for New Media
Opinion Space is a new interface that allows participants to visualize and navigate through a diversity of textual responses to a discussion question. New participants are first characterized by their stated opinions on five baseline statements, and this information is used with principal component analysis (PCA) to build a planar map. This visualization allows participants to quickly visualize the diversity of opinions and gives them more control over how to explore the ongoing discussion. To help participants cope with information overload, we are developing new spatial and statistical models for collaboratively identifying the most insightful responses. Specifically, we seek to highlight the responses that promote consensus across a diverse group of participants. We propose a model based on ensemble learning theory for weighting and aggregating response ratings that gives greater influence to positive ratings from users who have previously tended to disagree with the author of the response, and we compare it with more traditional methods. The model has the added benefit of introducing mechanisms that resist manipulation by false ratings and sybil attacks.
Figure 1: Opinion Space on the U.S. Department of State website.
- E. Bitton. A Spatial Model for Collaborative Filtering of Comments in an Online Discussion Forum. Proceedings of the ACM Conference on Recommender Systems, ACM 2009.
- T. Nathanson. Algorithms, Models, and Systems for Eigentaste-Based Collaborative Filtering and Visualization. Masters report, University of California at Berkeley.
- S. Faridani, E. Bitton, K. Goldberg, K. Ryokai. Opinion Space: A Scalable Tool for Browsing Online Comments. Proceedings of the 28th International ACM Conference on Human Factors in Computing Systems (CHI), 2010.