Interactive Visual Classification of Multivariate Data
Ke-Bing Zhang, Mehmet Ali Orgun, Rajan Shankaran, Du Zhang · 2012
This study proposes a visual approach for classification of multivariate data based on the enhanced separation feature of a visual technique, called Hypothesis-Oriented Verification and Validation by Visualization (HOV3). In this approach, the user first builds up a visual classifier from a training dataset based on its data projection plotted by HOV3with a statistical measurement of the training dataset on a 2d space where data points with the same class label are well grouped. Then the user classifies unlabeled data points by projecting them with the labeled data points of the visual classifier together in order to collect the unlabeled data points overlapped by the labeled ones. As a result, this study provides a method which is intuitive and easy to use for data classification by visualization.