The task delves into the groundbreaking computational models for processing information directly on graphs, Graph Neural Networks (GNNs). We were to find author and co-author suggestions based on the given dataset. The project has been documented in 3 parts based on the task given.The model and other files have been uploaded along with the documents.
This part consists the Exploratory Data Analysis of the dataset, using Cypher queries and methologies enlisted within the documentation. You can refer to the documentation in the uploaded files or from the link given below: (https://docs.google.com/document/d/14hXDhHROXqxIsnfMTE9XUoVQ5_bHN_0qB-upuCDtyM4/edit?usp=sharing)
In this section I use PyTorch Geometric and GNN model to predict 5 most likely authors from the graph, who could co-author with the given author. You can refer to the documentation in the uploaded files or from the link given below: (https://docs.google.com/document/d/10AJrBY4tYDY6kR3r_sNLUXEk7394B3RCUSZSMk4FbU4/edit?usp=sharing)
This phase of the task required to deploy the model to the cloud, I used AWS Lambda to perform the same. The API and other information have been given in the document. (https://docs.google.com/document/d/140vppz1Dx-ls8KdkFCNbQLAQdw7bDB9Y3Y7su66UoXA/edit?usp=sharing)
Thank you so much for your time!