Validating and Refining Clusters via Visual Rendering

Clustering is an important technique for understanding and analysis of large multi-dimensional datasets in many scientific applications. Most of clustering research to date has been focused on developing automatic clustering algorithms or cluster validation methods. The automatic algorithms are known to work well in dealing with clusters of regular shapes, e.g. compact spherical shapes, but may incur higher error rates when dealing with arbitrarily shaped clusters. Although some efforts have been devoted to addressing the problem of skewed datasets, the problem of handling clusters with irregular shapes is still in its infancy, especially in terms of dimensionality of the datasets and the precision of the clustering results considered. Not surprisingly, the statistical indices works ineffective in validating clusters of irregular shapes, too. In this paper, we address the problem of cluster rendering of skewed datasets by introducing a series of visual rendering techniques and a visual framework (VISTA). A main idea of the VISTA approach is to capitalize on the power of visualization and interactive feedbacks to encourage domain experts to participate in the clustering revision and clustering validation process. The VISTA system has two unique features. First, it implements a linear and reliable mapping model to visualize k-dimensional data sets in a 2D star-coordinate space. Second, it provides a rich set of userfriendly and yet effective interactive rendering operations, allowing users to validate and interactively refine the cluster structure based on their visual experience as well their domain knowledge.