{
  "id": 19983,
  "title": "driver ID?",
  "url": "/competitions/state-farm-distracted-driver-detection/discussion/19983",
  "author_name": "",
  "post_date": "2016-04-06T20:50:39.703Z",
  "votes": 12,
  "comment_count": 7,
  "views": 1642,
  "content": "<p>Could we have a data file containing the img_number to driver ID mapping for the train set? it will help us to do reliable local CV by splitting the train data based on drivers as well (as the private LB has different drivers from the train set).</p>",
  "messages": [
    {
      "id": "114013",
      "postDate": "04/06/2016 20:50:39",
      "content": "<p>Could we have a data file containing the img_number to driver ID mapping for the train set? it will help us to do reliable local CV by splitting the train data based on drivers as well (as the private LB has different drivers from the train set).</p>",
      "rawMarkdown": "Could we have a data file containing the img_number to driver ID mapping for the train set? it will help us to do reliable local CV by splitting the train data based on drivers as well (as the private LB has different drivers from the train set).",
      "votes": null
    },
    {
      "id": "114030",
      "postDate": "04/07/2016 04:27:02",
      "content": "<p>+1. Also is it possible to provide the &quot;frame number&quot; of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.</p>",
      "rawMarkdown": "1. Also is it possible to provide the \"frame number\" of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.",
      "votes": null
    },
    {
      "id": "114050",
      "postDate": "04/07/2016 08:28:59",
      "content": "<p>[quote=felixlaumon;114030]</p>\n\n<p>Also is it possible to provide the &quot;frame number&quot; of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.</p>\n\n<p>[/quote]</p>\n\n<p>That has been noted in the data description as removed on purpose. It is likely that in the controlled environment where the images were captured, that the timings are </p>\n\n<p>a) not natural or representative of the real-world application</p>\n\n<p>and</p>\n\n<p>b) correlated with the target class</p>",
      "rawMarkdown": "[quote=felixlaumon;114030]\r\n\r\nAlso is it possible to provide the \"frame number\" of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.\r\n\r\n[/quote]\r\n\r\nThat has been noted in the data description as removed on purpose. It is likely that in the controlled environment where the images were captured, that the timings are \r\n\r\n a) not natural or representative of the real-world application\r\n\r\nand\r\n\r\n b) correlated with the target class",
      "votes": null
    },
    {
      "id": "114153",
      "postDate": "04/07/2016 20:34:52",
      "content": "<p>@felixlaumon, we don't have that information, and we don't intend to give that information. </p>\n\n<p>@woshialex, good idea. I'll see what I can do to make the driver ids available. </p>",
      "rawMarkdown": "felixlaumon, we don't have that information, and we don't intend to give that information. \r\n\r\n@woshialex, good idea. I'll see what I can do to make the driver ids available.",
      "votes": null
    },
    {
      "id": "114159",
      "postDate": "04/07/2016 20:54:00",
      "content": "<p>driver_imgs_list.csv is uploaded to the <a href=\"https://www.kaggle.com/c/state-farm-distracted-driver-detection/data\">data</a> page. </p>",
      "rawMarkdown": "driver_imgs_list.csv is uploaded to the [data][1] page. \r\n\r\n\r\n  [1]: https://www.kaggle.com/c/state-farm-distracted-driver-detection/data",
      "votes": null
    },
    {
      "id": "114161",
      "postDate": "04/07/2016 21:18:43",
      "content": "<p>Thank you, Wendy.</p>\n\n<p>Wow, only 26 drivers? They do take a lot of pictures!</p>",
      "rawMarkdown": "Thank you, Wendy.\r\n\r\nWow, only 26 drivers? They do take a lot of pictures!",
      "votes": null
    },
    {
      "id": "114162",
      "postDate": "04/07/2016 21:24:04",
      "content": "<p>So the driver IDs of the test set are not going to be disclosed, are they?</p>",
      "rawMarkdown": "So the driver IDs of the test set are not going to be disclosed, are they?",
      "votes": null
    },
    {
      "id": "114169",
      "postDate": "04/07/2016 22:13:39",
      "content": "<p>@rcarson, correct. You're going to build something that predicts the classes without any knowledge about the driver. </p>",
      "rawMarkdown": "rcarson, correct. You're going to build something that predicts the classes without any knowledge about the driver.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 114030,
      "author_name": "felixlaumon",
      "author_url": "",
      "post_date": "04/07/2016 04:27:02",
      "content": "<p>+1. Also is it possible to provide the &quot;frame number&quot; of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114050,
      "author_name": "slobo777",
      "author_url": "",
      "post_date": "04/07/2016 08:28:59",
      "content": "<p>[quote=felixlaumon;114030]</p>\n\n<p>Also is it possible to provide the &quot;frame number&quot; of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.</p>\n\n<p>[/quote]</p>\n\n<p>That has been noted in the data description as removed on purpose. It is likely that in the controlled environment where the images were captured, that the timings are </p>\n\n<p>a) not natural or representative of the real-world application</p>\n\n<p>and</p>\n\n<p>b) correlated with the target class</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114153,
      "author_name": "wendykan",
      "author_url": "",
      "post_date": "04/07/2016 20:34:52",
      "content": "<p>@felixlaumon, we don't have that information, and we don't intend to give that information. </p>\n\n<p>@woshialex, good idea. I'll see what I can do to make the driver ids available. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114159,
      "author_name": "wendykan",
      "author_url": "",
      "post_date": "04/07/2016 20:54:00",
      "content": "<p>driver_imgs_list.csv is uploaded to the <a href=\"https://www.kaggle.com/c/state-farm-distracted-driver-detection/data\">data</a> page. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114161,
      "author_name": "jiweiliu",
      "author_url": "",
      "post_date": "04/07/2016 21:18:43",
      "content": "<p>Thank you, Wendy.</p>\n\n<p>Wow, only 26 drivers? They do take a lot of pictures!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114162,
      "author_name": "jiweiliu",
      "author_url": "",
      "post_date": "04/07/2016 21:24:04",
      "content": "<p>So the driver IDs of the test set are not going to be disclosed, are they?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 114169,
      "author_name": "wendykan",
      "author_url": "",
      "post_date": "04/07/2016 22:13:39",
      "content": "<p>@rcarson, correct. You're going to build something that predicts the classes without any knowledge about the driver. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "114013": "Could we have a data file containing the img_number to driver ID mapping for the train set? it will help us to do reliable local CV by splitting the train data based on drivers as well (as the private LB has different drivers from the train set).",
    "114030": "1. Also is it possible to provide the \"frame number\" of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.",
    "114050": "[quote=felixlaumon;114030]\r\n\r\nAlso is it possible to provide the \"frame number\" of the images? It should be a lot more interesting if we are allowed to leverage the temporal aspect of the dataset.\r\n\r\n[/quote]\r\n\r\nThat has been noted in the data description as removed on purpose. It is likely that in the controlled environment where the images were captured, that the timings are \r\n\r\n a) not natural or representative of the real-world application\r\n\r\nand\r\n\r\n b) correlated with the target class",
    "114153": "felixlaumon, we don't have that information, and we don't intend to give that information. \r\n\r\n@woshialex, good idea. I'll see what I can do to make the driver ids available.",
    "114159": "driver_imgs_list.csv is uploaded to the [data][1] page. \r\n\r\n\r\n  [1]: https://www.kaggle.com/c/state-farm-distracted-driver-detection/data",
    "114161": "Thank you, Wendy.\r\n\r\nWow, only 26 drivers? They do take a lot of pictures!",
    "114162": "So the driver IDs of the test set are not going to be disclosed, are they?",
    "114169": "rcarson, correct. You're going to build something that predicts the classes without any knowledge about the driver."
  },
  "source": "meta"
}