An Efficient Automatic Meta-Path Selection for Social Event Detection via Hyperbolic Space

Zitai Qiu, Congbo Ma, Jia Wu, Jian Yang

Keywords: Social Event Detection, Graph Neural Networks, Automatic Meta- Path, Hyperbolic Space

Abstract: Social events reflect changes in communities, such as natural disasters and emergencies. Timely detection of these situations can help residents and organizations in the community avoid danger and reduce losses. The complex nature of social messages makes social event detection on social media challenging. Existing methods usually construct social messages into heterogeneous information graphs to facilitate learning their semantic and structural information. However, they usually assume a fixed set of meta-paths, which often cannot describe real-world data sets well. On the other hand, a large number of social messages are not labeled due to expensive labeling work, which leads to an increase in model training costs. In order to solve the above challenges, we proposed a Heterogeneous Information Graph Hyperbolic space Automatic Meta-path selection model (GraphHAM), an efficient model that automatically selects meta-path and combines hyperbolic space to learn information on social media. In particular, we apply an efficient automatic meta-path selection technique and convert the selected meta-path into a vector. We then designed a novel hyperbolic MLP to further learn the semantic and structural information of social information. Extensive experiments show that GraphHAM can achieve outstanding performance on real-world data using only 20% of the whole dataset as the training set.

Track: Social Networks, Social Media, and Society

Submission Guidelines Scope: Yes

Submission Guidelines Blind: Yes

Submission Guidelines Format: Yes

Submission Guidelines Limit: Yes

Submission Guidelines Authorship: Yes

Student Author: Yes

Submission Number: 1265