Scheduled System Maintenance:
Some services will be unavailable Sunday, March 29th through Monday, March 30th. We apologize for the inconvenience.
By Topic

Interactive local clustering operations for high dimensional data in parallel coordinates

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

The purchase and pricing options are temporarily unavailable. Please try again later.
4 Author(s)
Peihong Guo ; Key Lab. of Machine Perception (Minist. of Educ.), Peking Univ., Beijing, China ; He Xiao ; Zuchao Wang ; Xiaoru Yuan

In this paper, we propose an approach of clustering data in parallel coordinates through interactive local operations. Different from many other methods in which clustering is globally applied to the whole dataset, our interactive scheme allows users to directly apply attractive and repulsive operators at regions of interests, taking advantages of an electricity interaction metaphor, for clutter reduction and cluster detection. Our design enables users to interact directly with the parallel coordinate plots and provides great flexibility in exploring and revealing underlying patterns. With instant feedback, our work allows users to dynamically adjust the clustering parameters to reach an optimum. We also supply the user with a graph indicating the logical relationship between clusters. Our experiments show that our scheme is more efficient than traditional methods in performing visual analysis tasks.

Published in:

Pacific Visualization Symposium (PacificVis), 2010 IEEE

Date of Conference:

2-5 March 2010