{
  "fetched_at": "2026-10-20T07:09:15Z",
  "source_url": "https://www.kaggle.com/competitions/bankruptcy-risk-prediction/data",
  "pages": [
    {
      "name": "data-description",
      "content": "This dataset is provided by one of the banks. Some of the variables are encoded due to confidentiality issues. That's why they are presented in conventional units. It doesen't affect the learning process at all, so feel free to participate.\n\nVariables are presented in categorical and numerical forms.\n\nTo make it easier to code, there are they, divided into these 2 categories:\n`term`, `sum`, `payment`, `residence_since`, `age`, `credits`, `dependents` are numerical variables.\n\n`status`, `credit_report`, `reason`, `savings`, `employment`, `marital_status`, `guarantees`, `estate`, `other_credits`, `accommodation`, `qualification`, `phone` and `immigrant` are categorical.\n\n`id` is the identification number of an example, `bancruptcy` is the target variable in the train set.\n\nVariables description is presented in the datasets variables description.\n\nThe game is on!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "This competition is about predicting person's bankrupcy. You'll have dataset with depersonalized values from one of the banks. Your aim is to predict is a person bankrupt or not.\n\nThe full dataset contains one thousand examples, and as always, splitted into `train` and `test` set with the size of 800 and 200 respectively.\nThis is a classification task and the purpose is to predict person's bankruptcy for the test set.\n\nYou'll have to provide your solution file, that contains predicted probabilities of bankruptcies. The probabilities are linked with id's of the examples, so there will be 2 columns in your solution files.\n\nShortly, metric is `ROC AUC`.\n\nMore about submission file format and competition metric is in `metric and submission ` tab on the left.\n\nDataset and variables description are presented in the `Data` tab.\n\nEnjoy the participation!\n\n> The hardest part of any deal is starting!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Metric And Submission",
      "content": "## Metric\nThis competition metric is `ROC AUC`(the measure of the ability of a classifier to distinguish between classes). Full form is Receiver Operating Characteristic, Area Under the Curve\n\nUsually, it is used in classification tasks evaluations\n\n## Submission Format\nSubmission files should contain two columns: `id` and `proba`\n`id` should contain example's identificator, `proba` have to be a float number of the probability of a true label\n\nAll files should have a `.csv` extension with standard pandas csv saving options\nSaving example is -\n```\ndf.to_csv('submission.csv', index=False)\n```\n\n## Format example:\nid,proba\n1,0.16270209934189291\n2,0.21975484280643935\n3,0.17125674517466864\n4,0.74107274730291206\n\nin csv submission file.\n\nExample of a submission file is presented in `Data` tab.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "There are no special rules for this competition.\n\nBut notice, cheating is not welcome.\n\n**Simple conditions:**\n\n— Maximum 2 people in one team\n\n— 5 submitions per day to prevent spamming\n\n\nFeel free to pin your notebooks, but remember, pinning notebooks means that everyone can read it.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}