Since I want to keep it simple, I will use the popular Zachary’s Karate Club graph dataset. Here, the nodes represent 34 students who were involved in the club and the links represent 78 different interactions between pairs of members outside the club. There are two different types of labels i.e, the two factions.

Node Classification: In this task, our aim is to build a model to predict the labels of the nodes i.e, the factions joined by the students.

We divide the graph into train and test sets (in the ratio 70:30) using a specific seed value. The same train and test splits will be used to build predictive models with different libraries so that we can have a comparison.

Once we have chosen the task, there are 6 simple steps to build a predictive model:

  1. Data preparation
  2. Build the custom dataset
  3. Choose a GNN method
  4. Train the model
  5. Perform hyperparameter tuning
  6. Predict the node labels on test data

All three python libraries have an implementation where we can transform our graph into the library’s custom dataset and use it to train the Graph Neural Network model.

Continue reading: https://towardsdatascience.com/lets-talk-about-graph-neural-network-python-libraries-a0b23ec983b0?source=rss—-7f60cf5620c9—4

Source: towardsdatascience.com