Novel directed hypergraph p-Laplacian based semi-supervised learning method: theory and algorithms

To deal with irregular data structure, graph-based semi-supervised learning methods have been developed to solve classification problems, especially node classification problems. However, most semi-supervised learning methods have been developed for the un-directed graph data model. This paper will present the novel semi-supervised learning methods for directed hypergraphs. In other words, we will develop the novel-directed hypergraph Laplacian-based semi-supervised learning methods in detail and the novel-directed hypergraph p-Laplacian-based semi-supervised learning method. These methods can then solve numerous node classification tasks for graph data models.