The diversity donut: enabling participant control over the diversity of recommended responses

Authors:
David Wong;Siamak Faridani;Ephrat Bitton;Björn Hartmann;Ken Goldberg
Affiliations:
University of California, Berkeley, Berkeley, CA, USA;University of California, Berkeley, Berkeley, CA, USA;University of California, Berkeley, Berkeley, CA, USA;University of California, Berkeley, Berkeley, CA, USA;University of California, Berkeley, Berkeley, CA, USA
Venue:
CHI '11 Extended Abstracts on Human Factors in Computing Systems
Year:
2011

Citing 4
Cited 1

Canonical Correlation Analysis: An Overview with Application to Learning Methods

Neural Computation
A spatial model for collaborative filtering of comments in an online discussion forum

Proceedings of the third ACM conference on Recommender systems
Opinion space: a scalable tool for browsing online comments

Proceedings of the SIGCHI Conference on Human Factors in Computing Systems
Presenting diverse political opinions: how and how much

Proceedings of the SIGCHI Conference on Human Factors in Computing Systems

Diversifying user comments on news articles

WISE'12 Proceedings of the 13th international conference on Web Information Systems Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

Most online discussion interfaces organize textual responses using linear lists. Such lists do not scale to the number of responses and cannot convey the diversity of the participants who have contributed. The Opinion Space system is designed to address these issues. In this paper, we augment Opinion Space with two features. The first is a new user interface tool and recommendation system: the Diversity Donut (Figure 1). While the Diversity Donut did not establish a statistical advantage over other recommendation methods, participant self-reported data suggested that participants found the Diversity Donut to yield the most diverse set of comments. The second contribution is a new dimensionality reduction technique in Opinion Space: Canonical Correlation Analysis (CCA). Our analysis suggests that CCA is a better algorithm for opinion visualization than Principal Component Analysis (PCA).