Visual to Parametric Interaction Algorithm

• Activity Number: 678
• Type: Topic Contributed
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Xinran Hu*+ and Scotland Charles Leman and Leanna House and Dipanya Maiti and Chris North and Lauren Bradel
• Companies: Virginia Tech and Virginia Tech and Virginia Tech and Virginia Tech and Virginia Tech and Virginia Tech
• Keywords: Visual Analytics ; Multidimensional Scaling ; Visualization

Abstract

When high-dimensional data is visualized in a 2D plane, by using a parametric projection algorithm, users may wish to make adjustments to the orientation of data objects to reflect their domain knowledge. However, few users are well-versed in the algorithms behind the visualization , making parameter tweaking more of a guessing game than a series of decisive interactions. Translating user interactions into algorithmic input is a key component of Visual to Parametric Interaction (V2PI). Instead of adjusting parameters, users move data objects on the screen, which then updates the underlying statistical model. However, we have found that some distinctly "unmoved" objects are as important as the "moved" objects in learning the user's intent. We are interested in designing an interpretation algorithm that specifically pays attention to these important "unmoved" objects.