When multiple column data are presented in graph.

[Jeriousman]

Let's say we have data features like [age, gender, score, income]. When we make node feature for this person in a hetero graph data, data['person'].x = ... # [num_persons, num_features_person] Do we normally concatenate [age, gender, score, income] for num_features_person? Is it how normally we do? If it is true the shape will be data['person'].x = [N, 4] I presume. Or should we make different features as a single feature in graph like data['person'].age, data['person'].gender, data['person'].score, data['person'].income?

It would be much appreciated to know common practice. Thank you.

[rusty1s]

data["person"].x should include all features of a person. As such, one concatenates all these different features into a single big feature matrix.

[Jeriousman]

Thank you for your great help!