{
  "id": 23405,
  "title": "Missing values distribution in Numeric train data",
  "url": "/competitions/bosch-production-line-performance/discussion/23405",
  "author_name": "",
  "post_date": "2016-09-06T12:16:42.997Z",
  "votes": 2,
  "comment_count": 6,
  "views": 1037,
  "content": "<p><a href=\"http://rpubs.com/saikumar/207032\">Understanding distribution of missing values</a></p>\n\n<p><a href=\"http://rpubs.com/saikumar/207035\">Understanding distribution of missing values for positive and negative classes</a></p>",
  "messages": [
    {
      "id": "134383",
      "postDate": "09/06/2016 12:16:42",
      "content": "<p><a href=\"http://rpubs.com/saikumar/207032\">Understanding distribution of missing values</a></p>\n\n<p><a href=\"http://rpubs.com/saikumar/207035\">Understanding distribution of missing values for positive and negative classes</a></p>",
      "rawMarkdown": "[Understanding distribution of missing values][1]\r\n\r\n\r\n[Understanding distribution of missing values for positive and negative classes][2]\r\n\r\n\r\n  [1]: http://rpubs.com/saikumar/207032\r\n  [2]: http://rpubs.com/saikumar/207035",
      "votes": null
    },
    {
      "id": "134396",
      "postDate": "09/06/2016 13:49:19",
      "content": "<p>What are the meanings of negative and positive classes and why they are corrolating?</p>",
      "rawMarkdown": "What are the meanings of negative and positive classes and why they are corrolating?",
      "votes": null
    },
    {
      "id": "134398",
      "postDate": "09/06/2016 14:00:22",
      "content": "<p>Can you please explain a little bit more about it?</p>",
      "rawMarkdown": "Can you please explain a little bit more about it?",
      "votes": null
    },
    {
      "id": "134422",
      "postDate": "09/06/2016 15:05:36",
      "content": "<p>Data is very sparse, and it is expected in any production line. \nX-Axis - Column name ,\nY-Axis - % of missing values</p>\n\n<hr>\n\n<p>In the first graph, I was trying to plot percentage of columns which are falling under certain threshold.</p>\n\n<h2>77% of columns have more than 80% missing values.</h2>\n\n<p>in the second graph, I wanted to see percentage of missing values in negative(0) and positive(1) classes.</p>\n\n<p>Both have similar distribution. </p>",
      "rawMarkdown": "Data is very sparse, and it is expected in any production line. \r\nX-Axis - Column name ,\r\nY-Axis - % of missing values\r\n\r\n----------\r\n\r\n\r\nIn the first graph, I was trying to plot percentage of columns which are falling under certain threshold.\r\n\r\n77% of columns have more than 80% missing values.\r\n----------\r\n in the second graph, I wanted to see percentage of missing values in negative(0) and positive(1) classes.\r\n\r\nBoth have similar distribution.",
      "votes": null
    },
    {
      "id": "134813",
      "postDate": "09/08/2016 16:09:16",
      "content": "<p>IMHO, the name &quot;missing&quot; or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation do not make sense.</p>",
      "rawMarkdown": "IMHO, the name \"missing\" or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation do not make sense.",
      "votes": null
    },
    {
      "id": "134893",
      "postDate": "09/09/2016 04:37:59",
      "content": "<p>[quote=Grzegorz Sionkowski;134813]</p>\n\n<p>IMHO, the name &quot;missing&quot; or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation may have no sense.</p>\n\n<p>[/quote]</p>\n\n<p>You are right!! </p>",
      "rawMarkdown": "[quote=Grzegorz Sionkowski;134813]\r\n\r\nIMHO, the name \"missing\" or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation may have no sense.\r\n\r\n[/quote]\r\n\r\nYou are right!!",
      "votes": null
    },
    {
      "id": "135171",
      "postDate": "09/11/2016 19:21:10",
      "content": "<p>Yeah Grzegorz, I agree with you totally: NA means a part did not go throw a certain LSF ( Line Station and Feature).  So it does not make sense to impute something that does not even exist. </p>",
      "rawMarkdown": "Yeah Grzegorz, I agree with you totally: NA means a part did not go throw a certain LSF ( Line Station and Feature).  So it does not make sense to impute something that does not even exist.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 134396,
      "author_name": "mojtabaebrahimi85",
      "author_url": "",
      "post_date": "09/06/2016 13:49:19",
      "content": "<p>What are the meanings of negative and positive classes and why they are corrolating?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134398,
      "author_name": "mrugankakarte",
      "author_url": "",
      "post_date": "09/06/2016 14:00:22",
      "content": "<p>Can you please explain a little bit more about it?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134422,
      "author_name": "saikumarallaka",
      "author_url": "",
      "post_date": "09/06/2016 15:05:36",
      "content": "<p>Data is very sparse, and it is expected in any production line. \nX-Axis - Column name ,\nY-Axis - % of missing values</p>\n\n<hr>\n\n<p>In the first graph, I was trying to plot percentage of columns which are falling under certain threshold.</p>\n\n<h2>77% of columns have more than 80% missing values.</h2>\n\n<p>in the second graph, I wanted to see percentage of missing values in negative(0) and positive(1) classes.</p>\n\n<p>Both have similar distribution. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134813,
      "author_name": "sionek",
      "author_url": "",
      "post_date": "09/08/2016 16:09:16",
      "content": "<p>IMHO, the name &quot;missing&quot; or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation do not make sense.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134893,
      "author_name": "saikumarallaka",
      "author_url": "",
      "post_date": "09/09/2016 04:37:59",
      "content": "<p>[quote=Grzegorz Sionkowski;134813]</p>\n\n<p>IMHO, the name &quot;missing&quot; or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation may have no sense.</p>\n\n<p>[/quote]</p>\n\n<p>You are right!! </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 135171,
      "author_name": "chambox",
      "author_url": "",
      "post_date": "09/11/2016 19:21:10",
      "content": "<p>Yeah Grzegorz, I agree with you totally: NA means a part did not go throw a certain LSF ( Line Station and Feature).  So it does not make sense to impute something that does not even exist. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "134383": "[Understanding distribution of missing values][1]\r\n\r\n\r\n[Understanding distribution of missing values for positive and negative classes][2]\r\n\r\n\r\n  [1]: http://rpubs.com/saikumar/207032\r\n  [2]: http://rpubs.com/saikumar/207035",
    "134396": "What are the meanings of negative and positive classes and why they are corrolating?",
    "134398": "Can you please explain a little bit more about it?",
    "134422": "Data is very sparse, and it is expected in any production line. \r\nX-Axis - Column name ,\r\nY-Axis - % of missing values\r\n\r\n----------\r\n\r\n\r\nIn the first graph, I was trying to plot percentage of columns which are falling under certain threshold.\r\n\r\n77% of columns have more than 80% missing values.\r\n----------\r\n in the second graph, I wanted to see percentage of missing values in negative(0) and positive(1) classes.\r\n\r\nBoth have similar distribution.",
    "134813": "IMHO, the name \"missing\" or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation do not make sense.",
    "134893": "[quote=Grzegorz Sionkowski;134813]\r\n\r\nIMHO, the name \"missing\" or treating the abbreviation NA as Not Available value are  wrong in this competition. The abbreviation NA is still good, but it should be treated as Not Applicable, so some actions, e.g. imputation may have no sense.\r\n\r\n[/quote]\r\n\r\nYou are right!!",
    "135171": "Yeah Grzegorz, I agree with you totally: NA means a part did not go throw a certain LSF ( Line Station and Feature).  So it does not make sense to impute something that does not even exist."
  },
  "source": "meta"
}