{
  "id": 2533,
  "title": "Not able to open training set 1 file",
  "url": "/competitions/MerckActivity/discussion/2533",
  "author_name": "",
  "post_date": "2012-09-01T14:08:09.827Z",
  "votes": null,
  "comment_count": 6,
  "views": 2962,
  "content": "<p>hi friends,</p>\r\n<p>I downloaded the file TrainingSet file in .7z format, it was 33.63 MB. Now when I extracted the file, I got 15 files with .csv format.</p>\r\n<p>The size of ACT1_competition_training.csv is around 703 MB. Does everybody have the same file size. Its very large. Is there something wrong or the data provided is that much large?</p>",
  "messages": [
    {
      "id": "13754",
      "postDate": "09/01/2012 14:08:09",
      "content": "<p>hi friends,</p>\r\n<p>I downloaded the file TrainingSet file in .7z format, it was 33.63 MB. Now when I extracted the file, I got 15 files with .csv format.</p>\r\n<p>The size of ACT1_competition_training.csv is around 703 MB. Does everybody have the same file size. Its very large. Is there something wrong or the data provided is that much large?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "13757",
      "postDate": "09/01/2012 15:08:37",
      "content": "<p>Yes, ACT1 is a large file. Here are all the training file sizes:</p>\r\n<p>ACT_</p>\r\n<p>1: 676 MB<br>\r\n2: 98 MB<br>\r\n3: 61 MB<br>\r\n4: 15 MB<br>\r\n5: 39 MB<br>\r\n6: 638 MB<br>\r\n7: 14 MB<br>\r\n8: 111 MB<br>\r\n9: 49 MB<br>\r\n10: 124 MB<br>\r\n11: 63 MB<br>\r\n12: 91 MB<br>\r\n13: 67 MB<br>\r\n14: 47 MB<br>\r\n15: 54 MB</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14085",
      "postDate": "09/07/2012 23:07:53",
      "content": "<p>I used Vim or gVim to open larger files.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14281",
      "postDate": "09/12/2012 20:09:26",
      "content": "<p>nowadays companies have gotten very smart combining several competitions in single competition.</p>\r\n<p>Like hewlett - I had to score 10 essays to get rank of 24th overall. Could have easily been 10 separate competitions</p>\r\n<p>At $4000 spent per dataset (in Merck), this is a real steal for the company. Plus the size of the datasets - if you want to center and scale sets 1 or 6 you will easily need 8GB&#43; RAM. hence wondering whether it is worth participating in this competition</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14288",
      "postDate": "09/13/2012 00:58:24",
      "content": "<p>It's true sets 1 and 6 are large. Since the evaluation metric puts equal weight on each activity however, I could see punting a little on those.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14290",
      "postDate": "09/13/2012 03:57:00",
      "content": "<p>I plan to do the same.</p>\r\n<p>But releasing too many datasets in a competition is good for the competition host but not for competitors. We are having to do a lot more work&nbsp; - solving a lot more problems! I do this for fun - Maybe kaggle should in future restrict number of datasets to\r\n 2 or 3 per competition. They would benefit too by getting to host more competitions while also benefitting us.</p>\r\n<p>&nbsp;</p>\r\n<p>[quote=Shea Parkes;14288]</p>\r\n<p>It's true sets 1 and 6 are large. Since the evaluation metric puts equal weight on each activity however, I could see punting a little on those.</p>\r\n<p>[/quote]</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "14293",
      "postDate": "09/13/2012 12:45:40",
      "content": "<p>I like the old programming axiom: There's no such thing as 2. There can be one, or there can be many. As soon as there is more one, I wrap an extra loop around all of my code. For this contest, I tossed in a couple switches to treat sets 1 &amp; 6 differently\r\n due to their size.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 13757,
      "author_name": "joycenv",
      "author_url": "",
      "post_date": "09/01/2012 15:08:37",
      "content": "<p>Yes, ACT1 is a large file. Here are all the training file sizes:</p>\r\n<p>ACT_</p>\r\n<p>1: 676 MB<br>\r\n2: 98 MB<br>\r\n3: 61 MB<br>\r\n4: 15 MB<br>\r\n5: 39 MB<br>\r\n6: 638 MB<br>\r\n7: 14 MB<br>\r\n8: 111 MB<br>\r\n9: 49 MB<br>\r\n10: 124 MB<br>\r\n11: 63 MB<br>\r\n12: 91 MB<br>\r\n13: 67 MB<br>\r\n14: 47 MB<br>\r\n15: 54 MB</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14085,
      "author_name": "",
      "author_url": "",
      "post_date": "09/07/2012 23:07:53",
      "content": "<p>I used Vim or gVim to open larger files.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14281,
      "author_name": "rkirana",
      "author_url": "",
      "post_date": "09/12/2012 20:09:26",
      "content": "<p>nowadays companies have gotten very smart combining several competitions in single competition.</p>\r\n<p>Like hewlett - I had to score 10 essays to get rank of 24th overall. Could have easily been 10 separate competitions</p>\r\n<p>At $4000 spent per dataset (in Merck), this is a real steal for the company. Plus the size of the datasets - if you want to center and scale sets 1 or 6 you will easily need 8GB&#43; RAM. hence wondering whether it is worth participating in this competition</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14288,
      "author_name": "biasvariance",
      "author_url": "",
      "post_date": "09/13/2012 00:58:24",
      "content": "<p>It's true sets 1 and 6 are large. Since the evaluation metric puts equal weight on each activity however, I could see punting a little on those.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14290,
      "author_name": "rkirana",
      "author_url": "",
      "post_date": "09/13/2012 03:57:00",
      "content": "<p>I plan to do the same.</p>\r\n<p>But releasing too many datasets in a competition is good for the competition host but not for competitors. We are having to do a lot more work&nbsp; - solving a lot more problems! I do this for fun - Maybe kaggle should in future restrict number of datasets to\r\n 2 or 3 per competition. They would benefit too by getting to host more competitions while also benefitting us.</p>\r\n<p>&nbsp;</p>\r\n<p>[quote=Shea Parkes;14288]</p>\r\n<p>It's true sets 1 and 6 are large. Since the evaluation metric puts equal weight on each activity however, I could see punting a little on those.</p>\r\n<p>[/quote]</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 14293,
      "author_name": "biasvariance",
      "author_url": "",
      "post_date": "09/13/2012 12:45:40",
      "content": "<p>I like the old programming axiom: There's no such thing as 2. There can be one, or there can be many. As soon as there is more one, I wrap an extra loop around all of my code. For this contest, I tossed in a couple switches to treat sets 1 &amp; 6 differently\r\n due to their size.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "13754": "",
    "13757": "",
    "14085": "",
    "14281": "",
    "14288": "",
    "14290": "",
    "14293": ""
  },
  "source": "meta"
}