Leveraging Meta-path Contexts for Classification in Heterogeneous Information Networks

• DOI: 10.48550/arxiv.2012.10024
• Publication Date: 2020-01-01
• Authors: Xiang Li; Danhao Ding; Ben Kao; Yizhou Sun; Nikos Mamoulis
• Keywords: Conch

A heterogeneous information network (HIN) has as vertices objects of different types and edges the relations between objects, which are also various types. We study problem classifying in HINs. Most existing methods perform poorly when given scarce labeled training sets, that improve classification accuracy under such scenarios often computationally expensive. To address these problems, we propose ConCH, a graph neural model. ConCH formulates multi-task learning combines semi-supervised with self-supervised to learn from both unlabeled data. employs meta-paths, sequences object capture semantic relationships objects. co-derives embeddings context via convolution. It uses attention mechanism fuse embeddings. conduct extensive experiments evaluate performance against other 15 methods. Our results show is an effective efficient method for HIN classification.