{
  "fetched_at": "2026-10-19T19:05:15Z",
  "source_url": "https://www.kaggle.com/competitions/cs3319-01-project-1/data",
  "pages": [
    {
      "name": "data-description",
      "content": "This dataset is a citation network and is a connected graph. This dataset has six attributes: x (feature of nodes), y(label of nodes, only those for nodes in training set or validation set is valid), edge_index (edges), num_classes (number of classes of nodes), train_mask and val_mask (mark nodes in training set or validation set).",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "NULL",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Your submission should have two rows. Here is an example:\nID 0 1 2 3\n0  0 1 1 0\nThe valid input is 0 1 1 0, which means [[0, 1], [1, 0]] in the form of edge_index.\nWe will provide a function to convert edge_index into submission.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "1. Don't cheat, including copying classmate's and any open source code.\n2. A reasonable method is more important than adjusting parameters.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "This project is about node classification and graph enhancement. In this project, you will attempt to enhance the expressive power of the graph by adding or removing a certain number of edges. We have pre trained multiple GNN models with fixed parameters to test the performance of the enhanced image you submitted. After the final submission, we will retrain the GNN model using the submitted images. Your final score will take into account both of the above factors. The dataset and code we provide is based on PYG package.\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}