Interactive Graph Construction for Graph-Based Semi-Supervised Learning

Semi-supervised learning (SSL) provides a way to improve the performance of prediction models (e.g., classifier) via usage unlabeled samples. An effective and widely used method is construct graph that describes relationship between labeled Practical experience indicates quality significantly affects model performance. In this paper, we present visual analysis interactively constructs high-quality for better particular, propose an interactive construction based on large margin principle. We have developed river visualization hybrid combines scatterplot, node-link diagram, bar chart convey label propagation graph-based SSL. Based understanding propagation, user can select regions interest inspect modify graph. conducted two case studies showcase how our facilitates exploitation samples improving