Using one-hot encoding instead of bag of words

[nguyenhongson1902]

Why did you use one-hot encoding later although you introduced bag of words technique in your article?

def encode_label(labels):
    label_encoder = LabelEncoder()
    labels = label_encoder.fit_transform(labels)
    labels = to_categorical(labels)
    return labels, label_encoder.classes_

labels_encoded, classes = encode_label(labels)

Link to your article: Training Graph Convolutional Networks on Node Classification Task