{
  "id": 57000,
  "title": "Difference Between Linear Regression and Gradient Boosting",
  "url": "/competitions/avito-demand-prediction/discussion/57000",
  "author_name": "",
  "post_date": "2018-05-17T23:14:18.279344Z",
  "votes": 12,
  "comment_count": 16,
  "views": 0,
  "content": "<p><img src=\"https://pbs.twimg.com/media/C29JevqXEAAYhMs.jpg\" alt=\"LRvsGB\"></p>",
  "messages": [
    {
      "id": "330053",
      "postDate": "05/17/2018 23:14:18",
      "content": "<p><img src=\"https://pbs.twimg.com/media/C29JevqXEAAYhMs.jpg\" alt=\"LRvsGB\"></p>",
      "rawMarkdown": "![LRvsGB][1]\n\n\n  [1]: https://pbs.twimg.com/media/C29JevqXEAAYhMs.jpg",
      "votes": null
    },
    {
      "id": "330078",
      "postDate": "05/18/2018 01:55:59",
      "content": "<p>nice picture</p>",
      "rawMarkdown": "nice picture",
      "votes": null
    },
    {
      "id": "330176",
      "postDate": "05/18/2018 08:22:07",
      "content": "<p>Congrats for \"Grandmaster\" title Bojan!!</p>\n\n<p>That 5 dots under your profile picture looks nice..</p>",
      "rawMarkdown": "Congrats for \"Grandmaster\" title Bojan!!\n\nThat 5 dots under your profile picture looks nice..",
      "votes": null
    },
    {
      "id": "330183",
      "postDate": "05/18/2018 08:36:32",
      "content": "<p>I actually clicked thinking you posted something somewhat valuable...</p>",
      "rawMarkdown": "I actually clicked thinking you posted something somewhat valuable...",
      "votes": null
    },
    {
      "id": "330277",
      "postDate": "05/18/2018 13:08:52",
      "content": "<p>Sounds like a first line from a big data haiku :-)</p>",
      "rawMarkdown": "Sounds like a first line from a big data haiku :-)",
      "votes": null
    },
    {
      "id": "330310",
      "postDate": "05/18/2018 14:47:36",
      "content": "<p>I’m much more into cheesy limericks:</p>\n\n<p>There once was a linear regression</p>\n\n<p>That had a machine learning obsession</p>\n\n<p>It descended one day</p>\n\n<p>In a stochastic way</p>\n\n<p>But fell short of the boosted correction</p>",
      "rawMarkdown": "I’m much more into cheesy limericks:\n\nThere once was a linear regression\n\nThat had a machine learning obsession\n\nIt descended one day\n\nIn a stochastic way\n\nBut fell short of the boosted correction",
      "votes": null
    },
    {
      "id": "330317",
      "postDate": "05/18/2018 15:00:55",
      "content": "<p>Why would anybody downvote you for that???</p>",
      "rawMarkdown": "Why would anybody downvote you for that???",
      "votes": null
    },
    {
      "id": "330322",
      "postDate": "05/18/2018 15:12:07",
      "content": "<p>Probably for the same reason(s) people downvote my other posts. I don’t want to spell it out. :)</p>",
      "rawMarkdown": "Probably for the same reason(s) people downvote my other posts. I don’t want to spell it out. :)",
      "votes": null
    },
    {
      "id": "330333",
      "postDate": "05/18/2018 15:36:33",
      "content": "<p>Apparently I can't insert an image due to 'trouble submitting your reply' - so here's a link to the perfect explanation: <a href=\"http://thinktheology.co.uk/blog/article/is_it_because_i_is_black\">http://thinktheology.co.uk/blog/article/is_it_because_i_is_black</a></p>",
      "rawMarkdown": "Apparently I can't insert an image due to 'trouble submitting your reply' - so here's a link to the perfect explanation: http://thinktheology.co.uk/blog/article/is_it_because_i_is_black",
      "votes": null
    },
    {
      "id": "330346",
      "postDate": "05/18/2018 16:16:00",
      "content": "<p>You nailed it!</p>",
      "rawMarkdown": "You nailed it!",
      "votes": null
    },
    {
      "id": "330364",
      "postDate": "05/18/2018 16:53:57",
      "content": "<p>Though sometimes this is the problem you need to solve...</p>\n\n<p><img src=\"http://i46.tinypic.com/wlr3tl.jpg\" alt=\"Though sometimes this is the problem you need to solve...\"></p>",
      "rawMarkdown": "Though sometimes this is the problem you need to solve...\n\n![Though sometimes this is the problem you need to solve...][1]\n\n\n  [1]: http://i46.tinypic.com/wlr3tl.jpg",
      "votes": null
    },
    {
      "id": "330419",
      "postDate": "05/18/2018 19:11:50",
      "content": "<p>nice depiction...</p>",
      "rawMarkdown": "nice depiction...",
      "votes": null
    },
    {
      "id": "333352",
      "postDate": "05/25/2018 01:32:39",
      "content": "<p>I've clicked on this post multiple times now with high hopes you've shared some great explanation. I don't know how I keep falling for your games. </p>",
      "rawMarkdown": "I've clicked on this post multiple times now with high hopes you've shared some great explanation. I don't know how I keep falling for your games.",
      "votes": null
    },
    {
      "id": "335475",
      "postDate": "05/29/2018 19:51:11",
      "content": "<p>I realize this is a light-hearted thread, but I wanted to share a link to an excellent PyData London talk entitled \"Winning with Simple, even Linear, Models\"  by Vincent D. Warmerda. There is some real gold in this talk, especially on time series. This is the only talk I can recall in recent memory that I watched in full and on 1x speed. 😃 </p>\n\n<p><a href=\"https://youtu.be/68ABAU_V8qI\">Winning with Simple, even Linear, Models</a></p>",
      "rawMarkdown": "I realize this is a light-hearted thread, but I wanted to share a link to an excellent PyData London talk entitled \"Winning with Simple, even Linear, Models\"  by Vincent D. Warmerda. There is some real gold in this talk, especially on time series. This is the only talk I can recall in recent memory that I watched in full and on 1x speed. 😃 \n\n[Winning with Simple, even Linear, Models][1]\n\n\n  [1]: https://youtu.be/68ABAU_V8qI",
      "votes": null
    },
    {
      "id": "335520",
      "postDate": "05/29/2018 21:22:19",
      "content": "<p>This is cool, thanks. I <em>think</em> I have a cool approach for making good linear models in this competition... I learned the technique from the Mercari competition where I think it was a bit of a sleeper... We'll see if it actually works, but I'd be excited to share after the competition is over.</p>",
      "rawMarkdown": "This is cool, thanks. I *think* I have a cool approach for making good linear models in this competition... I learned the technique from the Mercari competition where I think it was a bit of a sleeper... We'll see if it actually works, but I'd be excited to share after the competition is over.",
      "votes": null
    },
    {
      "id": "335532",
      "postDate": "05/29/2018 22:01:58",
      "content": "<p>The Vowpal Platypus cannot be contained...</p>",
      "rawMarkdown": "The Vowpal Platypus cannot be contained...",
      "votes": null
    },
    {
      "id": "335536",
      "postDate": "05/29/2018 22:05:53",
      "content": "<p><a href=\"/marketneutral\">@marketneutral</a> I wish. I've spent a bunch of hours trying to get VP/VW to work here but haven't been successful on that front. I'm thinking of something different here. I'm still going to keep trying on VP/VW though. :)</p>",
      "rawMarkdown": "marketneutral I wish. I've spent a bunch of hours trying to get VP/VW to work here but haven't been successful on that front. I'm thinking of something different here. I'm still going to keep trying on VP/VW though. :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 330078,
      "author_name": "liuhdsgoal",
      "author_url": "",
      "post_date": "05/18/2018 01:55:59",
      "content": "<p>nice picture</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 330176,
      "author_name": "prashantkikani",
      "author_url": "",
      "post_date": "05/18/2018 08:22:07",
      "content": "<p>Congrats for \"Grandmaster\" title Bojan!!</p>\n\n<p>That 5 dots under your profile picture looks nice..</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 330183,
      "author_name": "maxhalford",
      "author_url": "",
      "post_date": "05/18/2018 08:36:32",
      "content": "<p>I actually clicked thinking you posted something somewhat valuable...</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 330277,
      "author_name": "konradb",
      "author_url": "",
      "post_date": "05/18/2018 13:08:52",
      "content": "<p>Sounds like a first line from a big data haiku :-)</p>",
      "votes": null,
      "replies": [
        {
          "id": 330310,
          "author_name": "tunguz",
          "author_url": "",
          "post_date": "05/18/2018 14:47:36",
          "content": "<p>I’m much more into cheesy limericks:</p>\n\n<p>There once was a linear regression</p>\n\n<p>That had a machine learning obsession</p>\n\n<p>It descended one day</p>\n\n<p>In a stochastic way</p>\n\n<p>But fell short of the boosted correction</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 330317,
          "author_name": "konradb",
          "author_url": "",
          "post_date": "05/18/2018 15:00:55",
          "content": "<p>Why would anybody downvote you for that???</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 330322,
          "author_name": "tunguz",
          "author_url": "",
          "post_date": "05/18/2018 15:12:07",
          "content": "<p>Probably for the same reason(s) people downvote my other posts. I don’t want to spell it out. :)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 330333,
          "author_name": "konradb",
          "author_url": "",
          "post_date": "05/18/2018 15:36:33",
          "content": "<p>Apparently I can't insert an image due to 'trouble submitting your reply' - so here's a link to the perfect explanation: <a href=\"http://thinktheology.co.uk/blog/article/is_it_because_i_is_black\">http://thinktheology.co.uk/blog/article/is_it_because_i_is_black</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 330346,
          "author_name": "tunguz",
          "author_url": "",
          "post_date": "05/18/2018 16:16:00",
          "content": "<p>You nailed it!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 330364,
      "author_name": "marketneutral",
      "author_url": "",
      "post_date": "05/18/2018 16:53:57",
      "content": "<p>Though sometimes this is the problem you need to solve...</p>\n\n<p><img src=\"http://i46.tinypic.com/wlr3tl.jpg\" alt=\"Though sometimes this is the problem you need to solve...\"></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 330419,
      "author_name": "shubhamz950",
      "author_url": "",
      "post_date": "05/18/2018 19:11:50",
      "content": "<p>nice depiction...</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 333352,
      "author_name": "ryches",
      "author_url": "",
      "post_date": "05/25/2018 01:32:39",
      "content": "<p>I've clicked on this post multiple times now with high hopes you've shared some great explanation. I don't know how I keep falling for your games. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 335475,
      "author_name": "marketneutral",
      "author_url": "",
      "post_date": "05/29/2018 19:51:11",
      "content": "<p>I realize this is a light-hearted thread, but I wanted to share a link to an excellent PyData London talk entitled \"Winning with Simple, even Linear, Models\"  by Vincent D. Warmerda. There is some real gold in this talk, especially on time series. This is the only talk I can recall in recent memory that I watched in full and on 1x speed. 😃 </p>\n\n<p><a href=\"https://youtu.be/68ABAU_V8qI\">Winning with Simple, even Linear, Models</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 335520,
          "author_name": "peterhurford",
          "author_url": "",
          "post_date": "05/29/2018 21:22:19",
          "content": "<p>This is cool, thanks. I <em>think</em> I have a cool approach for making good linear models in this competition... I learned the technique from the Mercari competition where I think it was a bit of a sleeper... We'll see if it actually works, but I'd be excited to share after the competition is over.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 335532,
          "author_name": "marketneutral",
          "author_url": "",
          "post_date": "05/29/2018 22:01:58",
          "content": "<p>The Vowpal Platypus cannot be contained...</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 335536,
          "author_name": "peterhurford",
          "author_url": "",
          "post_date": "05/29/2018 22:05:53",
          "content": "<p><a href=\"/marketneutral\">@marketneutral</a> I wish. I've spent a bunch of hours trying to get VP/VW to work here but haven't been successful on that front. I'm thinking of something different here. I'm still going to keep trying on VP/VW though. :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "330053": "![LRvsGB][1]\n\n\n  [1]: https://pbs.twimg.com/media/C29JevqXEAAYhMs.jpg",
    "330078": "nice picture",
    "330176": "Congrats for \"Grandmaster\" title Bojan!!\n\nThat 5 dots under your profile picture looks nice..",
    "330183": "I actually clicked thinking you posted something somewhat valuable...",
    "330277": "Sounds like a first line from a big data haiku :-)",
    "330310": "I’m much more into cheesy limericks:\n\nThere once was a linear regression\n\nThat had a machine learning obsession\n\nIt descended one day\n\nIn a stochastic way\n\nBut fell short of the boosted correction",
    "330317": "Why would anybody downvote you for that???",
    "330322": "Probably for the same reason(s) people downvote my other posts. I don’t want to spell it out. :)",
    "330333": "Apparently I can't insert an image due to 'trouble submitting your reply' - so here's a link to the perfect explanation: http://thinktheology.co.uk/blog/article/is_it_because_i_is_black",
    "330346": "You nailed it!",
    "330364": "Though sometimes this is the problem you need to solve...\n\n![Though sometimes this is the problem you need to solve...][1]\n\n\n  [1]: http://i46.tinypic.com/wlr3tl.jpg",
    "330419": "nice depiction...",
    "333352": "I've clicked on this post multiple times now with high hopes you've shared some great explanation. I don't know how I keep falling for your games.",
    "335475": "I realize this is a light-hearted thread, but I wanted to share a link to an excellent PyData London talk entitled \"Winning with Simple, even Linear, Models\"  by Vincent D. Warmerda. There is some real gold in this talk, especially on time series. This is the only talk I can recall in recent memory that I watched in full and on 1x speed. 😃 \n\n[Winning with Simple, even Linear, Models][1]\n\n\n  [1]: https://youtu.be/68ABAU_V8qI",
    "335520": "This is cool, thanks. I *think* I have a cool approach for making good linear models in this competition... I learned the technique from the Mercari competition where I think it was a bit of a sleeper... We'll see if it actually works, but I'd be excited to share after the competition is over.",
    "335532": "The Vowpal Platypus cannot be contained...",
    "335536": "marketneutral I wish. I've spent a bunch of hours trying to get VP/VW to work here but haven't been successful on that front. I'm thinking of something different here. I'm still going to keep trying on VP/VW though. :)"
  },
  "source": "meta"
}