{
  "id": 21025,
  "title": "Test data assumptions",
  "url": "/competitions/state-farm-distracted-driver-detection/discussion/21025",
  "author_name": "",
  "post_date": "2016-05-17T20:00:13.577Z",
  "votes": null,
  "comment_count": 2,
  "views": 501,
  "content": "<p>I looked at some of the test data but not sure if its safe to make these assumptions. Can someone confirm if these are true?</p>\n\n<ul>\n<li>Each driver in test set will have at least one c0/safe driving image</li>\n<li>A given drivers dress remains unchanged across all images</li>\n</ul>\n\n<p>Thanks</p>",
  "messages": [
    {
      "id": "120369",
      "postDate": "05/17/2016 20:00:13",
      "content": "<p>I looked at some of the test data but not sure if its safe to make these assumptions. Can someone confirm if these are true?</p>\n\n<ul>\n<li>Each driver in test set will have at least one c0/safe driving image</li>\n<li>A given drivers dress remains unchanged across all images</li>\n</ul>\n\n<p>Thanks</p>",
      "rawMarkdown": "I looked at some of the test data but not sure if its safe to make these assumptions. Can someone confirm if these are true?\r\n\r\n- Each driver in test set will have at least one c0/safe driving image\r\n- A given drivers dress remains unchanged across all images\r\n\r\nThanks",
      "votes": null
    },
    {
      "id": "120371",
      "postDate": "05/17/2016 20:39:36",
      "content": "<p>This looks true for the training data as well. I'd imagine that they generated the data the same way for each set. Or rather they randomly divided it. </p>\n\n<p>In other words, I'm with you in making those assumptions. </p>",
      "rawMarkdown": "This looks true for the training data as well. I'd imagine that they generated the data the same way for each set. Or rather they randomly divided it. \r\n\r\nIn other words, I'm with you in making those assumptions.",
      "votes": null
    },
    {
      "id": "120391",
      "postDate": "05/17/2016 23:34:14",
      "content": "<p>You can find more detail on how the data was generated <a href=\"https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/20162/mislabeled-training-images/115702#post115702\">here</a>. </p>",
      "rawMarkdown": "You can find more detail on how the data was generated [here][1]. \r\n\r\n\r\n  [1]: https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/20162/mislabeled-training-images/115702#post115702",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 120371,
      "author_name": "florianm",
      "author_url": "",
      "post_date": "05/17/2016 20:39:36",
      "content": "<p>This looks true for the training data as well. I'd imagine that they generated the data the same way for each set. Or rather they randomly divided it. </p>\n\n<p>In other words, I'm with you in making those assumptions. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 120391,
      "author_name": "spammy",
      "author_url": "",
      "post_date": "05/17/2016 23:34:14",
      "content": "<p>You can find more detail on how the data was generated <a href=\"https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/20162/mislabeled-training-images/115702#post115702\">here</a>. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "120369": "I looked at some of the test data but not sure if its safe to make these assumptions. Can someone confirm if these are true?\r\n\r\n- Each driver in test set will have at least one c0/safe driving image\r\n- A given drivers dress remains unchanged across all images\r\n\r\nThanks",
    "120371": "This looks true for the training data as well. I'd imagine that they generated the data the same way for each set. Or rather they randomly divided it. \r\n\r\nIn other words, I'm with you in making those assumptions.",
    "120391": "You can find more detail on how the data was generated [here][1]. \r\n\r\n\r\n  [1]: https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/20162/mislabeled-training-images/115702#post115702"
  },
  "source": "meta"
}