HIN_DRL: A random walk based dynamic network representation learning method for heterogeneous information networks

作者：

Highlights：

• Semantic-enriched dynamic heterogeneous network embedding framework.

• Automatically searching and extending meta-paths to extract semantics.

• Conducting edge-based dynamic random walk guided by extended meta-paths.

• Measuring transition probability by network structure, timestamp and semantics.

• Two strategies to optimize the quantity and length of node sequences dynamically.

摘要

•Semantic-enriched dynamic heterogeneous network embedding framework.•Automatically searching and extending meta-paths to extract semantics.•Conducting edge-based dynamic random walk guided by extended meta-paths.•Measuring transition probability by network structure, timestamp and semantics.•Two strategies to optimize the quantity and length of node sequences dynamically.

论文关键词：Dynamic representation learning,Heterogeneous information networks,Meta path,Dynamic random walk

论文评审过程：Received 13 June 2019, Revised 30 March 2020, Accepted 31 March 2020, Available online 23 May 2020, Version of Record 23 May 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113427