Journal
ELECTRONICS
Volume 12, Issue 2, Pages -Publisher
MDPI
DOI: 10.3390/electronics12020398
Keywords
graph neural networks; heterogeneous network; impact measurement; graph autoencoder
Ask authors/readers for more resources
Objectively evaluating representative papers in a specific scientific research field is important for the development of academia and scientific research institutions. The proposed GAEPIM framework uses a graph autoencoder based on heterogeneous networks to rank papers and find the most representative ones and their scientific institutions. The method outperforms several widely used baseline methods.
Objectively evaluating representative papers in a specific scientific research field is of great significance to the development of academia and scientific research institutions. Representative papers on achievements in scientific research can reflect the academic level and research characteristics of researchers and research institutions. The existing research methods are mainly based on external feature indicators and citation analysis methods, and the method of combining artificial intelligence is in its infancy. From the perspective of scientific research institutions, this paper proposes a graph autoencoder framework based on heterogeneous networks for the measurement of paper impact, named GAEPIM. Specifically, we propose two versions of GAEPIM based on a graph convolutional network and graph transformer network. The models rank papers in a specific research field and find the most representative papers and their scientific institutions. The proposed framework constructs a heterogeneous network of papers, institutions, and venues and simultaneously analyzes the semantic information of papers and the heterogeneous network structural information. Finally, based on the complex network information diffusion model, the proposed method performs better than several widely used baseline methods.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available