Skip to content

Utilizing graphical neural networks and embeddings on a medical database KEGG to perform link predictions and drug similarity systems.

License

Notifications You must be signed in to change notification settings

prakhargurawa/Drug-Similarity-and-Link-Prediction-using-Graph-Embeddings-on-Medical-Knowledge-Graph

Repository files navigation

Drug-Similarity-and-Link-Prediction-using-Graph-Embeddings-on-Medical-Knowledge-Graph

Utilizing graphical neural networks and embeddings on a medical database KEGG to perform link predictions and drug similarity system.

Section of KEGG dataset visualized as knowledge graph (using NetworkX)

Different color represent different entity present in the database that are Drugs, Diseases, Gene, Network and Pathway: alt text

Graph Neural Network pipeline to generate graphical embeddings, link prediction and drug similarity

alt text

Graph embedding visualized in 2D space using PCA

alt text

Comparision of different graphical embedding models on link prediction tasks

alt text

About

Utilizing graphical neural networks and embeddings on a medical database KEGG to perform link predictions and drug similarity systems.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published