{
  "fetched_at": "2026-10-20T00:30:33Z",
  "source_url": "https://www.kaggle.com/competitions/ece460j-fall24/data",
  "pages": [
    {
      "name": "Evaluation",
      "content": "Submissions are evaluated using the multi-class logarithmic loss. Each `id` in the test set had a single true class label, `Status`. For each `id`, you must submit a set of predicted probabilities for each of the three possible outcomes, e.g., `Status_C`, `Status_CL`, and `Status_D`.\n\nThe metric is calculated\n\n$$log loss = -\\frac{1}{N}\\sum_{i=1}^N\\sum_{j=1}^My_{ij}\\log(p_{ij}),$$\n\nwhere \\\\(N\\\\) is the number of rows in the test set, \\\\(M\\\\) is the number of outcomes (i.e., 3),  \\\\(log\\\\) is the natural logarithm, \\\\(y_{ij}\\\\) is 1 if row \\\\(i\\\\) has the ground truth label \\\\(j\\\\) and 0 otherwise, and \\\\(p_{ij}\\\\) is the predicted probability that observation \\\\(i\\\\) belongs to class \\\\(j\\\\).\n\nThe submitted probabilities for a given row are not required to sum to one because they are rescaled prior to being scored (each row is divided by the row sum). In order to avoid the extremes of the log function, predicted probabilities are replaced with \\\\(max(min(p,1-10^{-15}),10^{-15})\\\\).\n\n## Submission File\nFor each `id` row in the test set, you must predict probabilities of the three outcomes `Status_C`, `Status_CL`, and `Status_D` . The file should contain a header and have the following format:\n\n    id,Status_C,Status_CL,Status_D\n    7905,0.628084,0.034788,0.337128\n    7906,0.628084,0.034788,0.337128\n    7907,0.628084,0.034788,0.337128\n    etc.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "\nNo collaboration whatsoever is allowed in this competition.\nYou cannot discuss your solutions with anyone in or out of the class.\nYou can, however, use anything that is on-line and static -- so you can see what others have tried for other Kaggle competitions, and so on.\n\nIn summary….\n - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "data-description",
      "content": "The dataset for this competition (both train and test) was generated from a deep learning model trained on the [Cirrhosis Patient Survival Prediction](https://www.kaggle.com/datasets/joebeachcapital/cirrhosis-patient-survival-prediction) dataset. Feature distributions are close to, but not exactly the same, as the original. Feel free to use the original dataset as part of this competition, both to explore differences as well as to see whether incorporating the original in training improves model performance.\n\n## Files\n\n*   **train.csv** - the training dataset; `Status` is the categorical target; `C` (censored) indicates the patient was alive at `N_Days`, `CL` indicates the patient was alive at `N_Days` due to liver a transplant, and `D` indicates the patient was deceased at `N_Days`.\n*   **test.csv** - the test dataset; your objective is to predict the probability of each of the three `Status` values, e.g., `Status_C`, `Status_CL`, `Status_D`.\n*   **sample_submission.csv** - a sample submission file in the correct format",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Welcome to the DS Lab Kaggle competion!\n\nDo your best to make the best predictions by the closing time of the competition.\n\nAfter it closes, you will write a report that indicates what you tried, what worked and what did not.\nYou will be scored based on your private and public leaderboard score but also on your creativity on trying various ideas.\n\nyour task is to use a multi-class approach to predict the the outcomes of patients with cirrhosis. Good luck! The meanings of the features were not revealed by the data provider for privacy reasons.\n\nThe data may contain a few tricks. Explore and describe what you find.\n\nYou will get 5 submissions per day which resets at midnight UTC (7PM CT).",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}