Title:
PIVE: A Per-Iteration Visualization Environment for Supporting Real-time Interactions with Computational Methods
PIVE: A Per-Iteration Visualization Environment for Supporting Real-time Interactions with Computational Methods
Author(s)
Choo, Jaegul
Lee, Changhyun
Park, Haesun
Lee, Changhyun
Park, Haesun
Advisor(s)
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
Visual analytics has been gaining increasing interest due to its fascinating characteristic that leverages both humans’ visual perception and the power of computing. Although various computational methods are being proposed, they do not properly
support visual analytics. One of the biggest obstacles towards their real-time visual analytic integration is their high computational
complexity. As a way to tackle this problem, this paper presents PIVE, a Per-Iteration Visualization Environment for supporting
real-time interactive visualization with computational methods. The main idea behind PIVE is that most advanced computational
methods work by refining the solution iteratively. By visually delivering the result from each iteration to users, the proposed framework
enables users to quickly acquire the information that the computational method provides as well as the ability to perform continuous
interactions with them in real time. We show the effectiveness of PIVE in terms of real-time visualization and interaction capabilities by
customizing various dimension reduction methods such as principal component analysis, multidimensional scaling, and t-distributed
stochastic neighborhood embedding, and clustering method
s such as k-means and latent Dirichlet allocation.
Sponsor
Date Issued
2013
Extent
Resource Type
Text
Resource Subtype
Technical Report