{
  "id": 153616,
  "title": "Prediction Output Format",
  "url": "/competitions/alaska2-image-steganalysis/discussion/153616",
  "author_name": "",
  "post_date": "2020-05-25T15:47:26.008533Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Slightly confused about what the output format should be. This seems like a binary classification exercise where we need to predict values between <code>0</code> and <code>1</code>, where values closer to <code>0</code> mean the image does not contain stegonography and closer to <code>1</code> means it does.</p>\n\n<p>However, the <code>Overview</code> tab shows the following example under <code>Evaluation</code>:</p>\n\n<p><code>\nId,Label\n0001.jpg,0.1\n0002.jpg,0.99\n0003.jpg,1.2\n0004.jpg,-2.2\netc.\n</code></p>\n\n<p>How can a label be <code>1.1</code>, and more confusingly <code>-2.2</code>?</p>",
  "messages": [
    {
      "id": "860809",
      "postDate": "05/25/2020 15:47:26",
      "content": "<p>Slightly confused about what the output format should be. This seems like a binary classification exercise where we need to predict values between <code>0</code> and <code>1</code>, where values closer to <code>0</code> mean the image does not contain stegonography and closer to <code>1</code> means it does.</p>\n\n<p>However, the <code>Overview</code> tab shows the following example under <code>Evaluation</code>:</p>\n\n<p><code>\nId,Label\n0001.jpg,0.1\n0002.jpg,0.99\n0003.jpg,1.2\n0004.jpg,-2.2\netc.\n</code></p>\n\n<p>How can a label be <code>1.1</code>, and more confusingly <code>-2.2</code>?</p>",
      "rawMarkdown": "Slightly confused about what the output format should be. This seems like a binary classification exercise where we need to predict values between `0` and `1`, where values closer to `0` mean the image does not contain stegonography and closer to `1` means it does.\n\nHowever, the `Overview` tab shows the following example under `Evaluation`:\n\n```\nId,Label\n0001.jpg,0.1\n0002.jpg,0.99\n0003.jpg,1.2\n0004.jpg,-2.2\netc.\n```\n\nHow can a label be `1.1`, and more confusingly `-2.2`?",
      "votes": null
    },
    {
      "id": "861163",
      "postDate": "05/25/2020 23:07:41",
      "content": "<p>Check <a href=\"https://www.kaggle.com/c/alaska2-image-steganalysis/discussion/150592\">here</a></p>",
      "rawMarkdown": "Check [here](https://www.kaggle.com/c/alaska2-image-steganalysis/discussion/150592)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 861163,
      "author_name": "wuliaokaola",
      "author_url": "",
      "post_date": "05/25/2020 23:07:41",
      "content": "<p>Check <a href=\"https://www.kaggle.com/c/alaska2-image-steganalysis/discussion/150592\">here</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "860809": "Slightly confused about what the output format should be. This seems like a binary classification exercise where we need to predict values between `0` and `1`, where values closer to `0` mean the image does not contain stegonography and closer to `1` means it does.\n\nHowever, the `Overview` tab shows the following example under `Evaluation`:\n\n```\nId,Label\n0001.jpg,0.1\n0002.jpg,0.99\n0003.jpg,1.2\n0004.jpg,-2.2\netc.\n```\n\nHow can a label be `1.1`, and more confusingly `-2.2`?",
    "861163": "Check [here](https://www.kaggle.com/c/alaska2-image-steganalysis/discussion/150592)"
  },
  "source": "meta"
}