Abstract:
Graph link prediction is a classic graph analytical problem with many applications. Current link prediction methods use deep learning to compute features from subgraphs centered at two neighboring nodes and predict the link label between them. This formalism turns link prediction into graph classification. Graph pooling layers in the deep learning model lose information to extract fixed-size features for classification. We propose using graph theory line graphs to overcome this key limitation. Line graphs have unique edges for each node. Thus, link prediction problems in the original graph can be solved as node classification problems in its line graph instead of graph classification tasks. Our method outperforms state-of-the-art methods on fourteen datasets from different applications while having fewer parameters and high training efficiency.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
